DE10020178B4 - Apparatus and method for image binarization and apparatus and method for image acquisition - Google Patents

Abstract

Image binarizing apparatus comprising:
a block dividing unit that divides a multi-valued image into blocks;
a luminance value output unit that outputs luminance values for each pixel representing the multi-valued image;
a binarization threshold setting unit that sets a binarization threshold to be used when binarizing the multi-valued image;
a binarization unit that binarizes the multi-valued image based on the binarization threshold;
a low luminance threshold setting unit that sets a low luminance threshold to be used when removing low luminance values;
a block selecting unit that selects blocks to be processed from the blocks generated by the block dividing unit whose multivalued images are to be binarized by the binarizing unit;
a low luminance value removal unit that receives, from the luminance value output unit, the luminance values of those pixels that make up the selected block or blocks that removes the luminance values lower than the low luminance threshold value set by the adjustment unit for the luminance value low ...

Description

The The present invention relates to an apparatus for image binarization (Image digitization), an image capture apparatus, a process for picture-making (Image digitization), a method for image acquisition and a computer program for image capture and image binarization (image digitization). In particular, the invention relates to a technology in which a Picture binärisiert (digitized) becomes after shadows and irregularities have been removed in terms of brightness of a multi-valued image, wherein for input an image input device was used, the one does not have a constant light wave.

traditionally, is used to save a document as an electronic document generally a scanner, camera or scanner section a fax machine used as an image input device. In this type of scanner (scanner section) a light source is provided inside the apparatus and light, that is emitted by the light source and that of the document (original) is read by a CCD or the like. If necessary, the read image is then binarized (digitized) and saved. Because the characteristics of the changed light source and the optical system are constant, shades and Irregularities that in terms of brightness, also constant and can Therefore be easily corrected. Thus, high quality images can be output, and pictures can be easy and with high quality digitized or binarized become. The term "binarize" is used herein The term "digitize" is used predominantly to clarify that the present invention in particular the digitization in binary form or binary representation concerns. The "digitization" is therefore referred to in particular as "binarization". The The invention relates in particular to the reduction of a multi-valued pixel to a bivalent pixel.

On The other side has been in recent years with the increase in popularity of video cameras and digital cameras the need after that, it's possible should be to recognize characters in pictures of this Type of devices are entered. In particular with regard to The digital camera has begun to be used in various applications to use as portable information gathering tools, as the Number of pixels has increased significantly and their size decreases Has. For example, characters provide information, such as Documents or pamphlets, notepads or notes and Advertising materials or advertising leaflets sufficient information as a binary Pictures, and above In addition, the amount of disk space required is save this information, less than with multivalued images, so that it is beneficial to have the information binarized or digital image to save. Furthermore, digitized images can be transmitted Be by a fax machine is used, or they can be reused after undergoing a character recognition process were.

The Japanese Laid-Open Patent Application No. 3-237571 "Device For Calculation Binarized Threshold of Image "(" Device for computing digitized thresholds of an image ") discloses a technology to suitably digitize these types of digital images. According to this technology A high quality digitized image is obtained by using a digitizing circuit is provided, which is a digitization process with the image in a window, by taking the difference between the brightness of each pixel in the Window and the brightness of a specific pixel with a parameter which is proportional to the contrast of the observed image portion and by providing a determination circuit that the suitability of the binary or digital pattern determined by the digitizing circuit as an image pattern of a contour portion was obtained.

The Japanese Patent Application Laid-Open No. 7-212591 "Image Binarizing Device "also discloses a technology for digitization or binarization. According to this Technology becomes a histogram of nominal values from a multi-valued one Image, and the values that represent white pixels, and the Values representing black pixels are taken from the histogram certainly. A binarization or digitization threshold is then determined from the mean of it, and the multivalued image is binarized or digitized, based on the binarization threshold. This patent application also discloses a technology in the one picture in blocks is subdivided, the binarization threshold or the value, the white pixels or the value representing black pixels becomes for each Block determines blocks, that contain no characters are interpolated from the blocks in the environment, and a block threshold is used as the threshold for each Pixels accepted or accepted, which makes it possible that will be an unevenness and eliminates a shadow in terms of brightness in the image or removed. Thus, a binarization or digitization with high quality carried out using the technology described in this patent application is disclosed.

however The problems described below exist with the conventional one Technology. Namely the number, position and strength the light sources are different, and therefore shadows occur and brightness irregularities slightly in the photographed image because beyond the shadows and the brightness unevenness are not constant, it is not possible a uniform correction as in the scanner section of a photocopy machine or the like. Accordingly, have pictures taken using a digital camera have been photographed, the problem is that a binarization digitalization with high quality is not possible.

Even when pictures are photographed using a homogeneous light source will be about it In addition, a light reduction in the periphery by the properties of the optical system, resulting in the problem that a uniform Correction and a binarization high quality can not be achieved.

Furthermore For example, for a notebook or note, some are characters or groups of characters very small. If such a sticky note is photographed, there is a need, partly a binarization high quality perform.

at Japanese Laid-Open Patent Application No. 3-237571 "Device for Calculation Binaryized Threshold of Image "(" Device for computing a binarization threshold of a Picture ") It is required, the binarization circuit or Provide digitization circuit to image processing with each pixel in a window and provide the determining circuit, a pattern matching or pattern matching with outline sections as a picture pattern. Accordingly, there is the problem that the processing load has significantly increased and the processing speed consequently has decreased.

Furthermore It is disclosed in Japanese Patent Application Laid-Open No. 7-212591 "Image Binarizing Device "(" Bildbinärisierungs device ") in addition to Creation of a luminance value histogram necessary for all pixels, values to determine the white one and represent black pixels by loop processing is used, and it is necessary to improve processing the edges or edges perform, around the edges or edges to improve the characters and diagrams. Accordingly exists the problem that considerable Processing resources needed and the power consumption is just as high.

task The invention is a binarization or digitizing high quality with a multi-valued image perform.

above Task is through the objects the independent one claims solved. Advantageous developments emerge from the dependent claims.

Advantageous becomes a binarization high quality with a multi-valued image fast and with a low power consumption carried out.

Advantageous removed according to one Aspect of the present invention, the image binarizing apparatus the low luminance values of an object block on the base a mean luminance value of the surrounding blocks and sets the binarization threshold of the Object blocks based on the mean of the luminance values from which the low luminance values have been removed.

Advantageous removed according to one another aspect of the image binarization apparatus the low luminance values based on the average luminance values of the blocks in the environment, sets the binarization thresholds the object blocks based on the mean of the luminance values, of which the low ones Luminance values have been removed, and set the binarization thresholds, which are to be applied to each pixel of interpolation blocks on which Basis of the binarization thresholds from adjacent blocks of objects firmly.

Advantageous remove the image binarization apparatus According to yet another aspect, the low luminance values of object blocks based on the mean luminance values of the blocks in the environment and rounds out the mean of the luminance values, of which the low ones Luminance values have been removed to values within a predetermined range Range. These values then form the basis on which the binarization thresholds the object blocks be determined.

In yet another aspect, the image binarizing apparatus advantageously removes the low luminance values of the object blocks based on the mean luminance values of the blocks in the environment, and rounds the average of the luminance values from which the low luminance values have been removed to values within a predetermined range. These values then form the basis on which the binarization thresholds of the object blocks are determined. The binarization thresholds that appear on each pixel of the Inter poling values are set based on the binarization thresholds of adjacent object blocks.

According to one more Another aspect is the binarization thresholds of generated blocks on the basis of smoothed photometric values of generated screens or measuring ranges. ("Screens") generated.

According to one more In another aspect, the image capture apparatus advantageously sets the binarization thresholds. which are applied to every pixel of an interpolation block, on the basis of the smoothed photometric values of adjacent generated screens or measuring ranges.

Advantageous, according to one more In another aspect, the image capture apparatus completes the photometric Values of generated screens or measuring ranges to values within one predetermined range and sets the binarization thresholds generated blocks based on these values.

Advantageous, according to one more In another aspect, the image capture apparatus completes the photometric Values of adjacent generated screens or measuring ranges to values within a predetermined range and sets the binarization thresholds, which are applied to each pixel, based on an interpolation block on these values.

Advantageous removed according to one yet another aspect of the image binarization process the low luminance values of the object blocks based on the mean Luminance values of the blocks in the neighborhood. The binarization thresholds of object blocks are then based on the mean of the luminance values from which the low luminance values have been removed.

Advantageous removed according to one yet another aspect of the image binarization process the low luminance values of the object blocks based on the mean Luminance values of the blocks in the neighborhood. The binarization thresholds of object blocks are then based on the mean of the luminance values from which the luminance values have been removed. The binarization thresholds, which are to be applied to each pixel of the interpolation block then become based on the binarization thresholds of adjacent object blocks established.

The Image binarize according to one more Another aspect of the present invention removes the low Luminance values based on the mean luminance values of Blocks in the environment. The mean of the luminance values, of which the low Luminance values have been removed, then to values within of a predetermined range, and the binarization thresholds the object blocks are then determined based on these values.

According to one more In another aspect, the image binarization method removes the low ones Luminance values based on the mean luminance values of the blocks in the neighborhood. The mean values of the luminance values, of which the low luminance values have been removed then to values rounded within a predetermined range, and the binarization thresholds the object blocks are then determined based on these values. The binarization thresholds, which are applied to each pixel of the interpolation blocks then based on the binarization thresholds of adjacent ones object blocks established.

According to one more In another aspect, the image capture method sets binarization thresholds more generated blocks on the basis of smoothed photometric values of a generated screen or measuring range.

According to one more In another aspect, the image acquisition method sets the binarization thresholds that be applied to each pixel of an interpolation block the basis of smoothed photometric values of neighboring generated screens or measuring ranges firmly.

According to one more In another aspect, the image acquisition process completes the photometric values generated screens or measuring ranges to values within a predetermined Range and sets the binarization thresholds of generated blocks based on these values.

According to one more In another aspect, the image acquisition process completes the photometric values adjacent generated screens to values within a predetermined range Range and sets the binarization thresholds, which are applied to each pixel of the interpolation blocks the basis of these values.

According to one more yet another aspect will be beneficial a program and a computer readable Recording medium on which the program is stored to Computer to enable the functions of each step the image binarization process and the image acquisition method described above.

Further Features of the invention will be disclosed in the following description. Various features of different embodiments may be used together be combined.

1 Fig. 10 is a block diagram showing an example of a structure of an image binarizing apparatus of the first embodiment of the present invention when used in a digital camera.

2 Fig. 10 is a conceptual view of a light receiving portion of a CCD in a digital camera of the first embodiment.

3A and 3B For example, examples of divisions of multi-value images photographed using a digital camera of the first embodiment are shown in blocks.

4A and 4B show examples of sampling intervals for sampling pixels within a block.

5 FIG. 10 is a block diagram showing an example of a construction of a mean luminance value calculating means of a digital camera of the first embodiment. FIG.

6 Fig. 10 is a flow chart showing the flow of image data processing as much as the binarization of a multi-valued image in a digital camera of the first embodiment.

7 Fig. 10 is a block diagram showing an example of the apparatus construction when an image binarizing apparatus of the second embodiment of the present invention is used in a digital camera.

8th Fig. 10 is a block diagram showing an example of the apparatus construction when an image binarizing apparatus of the third embodiment of the present invention is used in a digital camera.

9A and 9B Fig. 10 are explanatory diagrams summarizing the calculation of binarization thresholds applied to each pixel within an interpolation block of a digital camera of the third embodiment.

10 Fig. 10 is a flowchart showing the flow of image data processing in terms of binarizing a multi-valued image in a digital camera of the third embodiment.

11 Fig. 10 is a block diagram showing an example of the apparatus construction of a digital camera in which binarization thresholds applied to each pixel are calculated and image binarization is performed for each pixel by using a CMOS sensor in the image input section ,

12 Fig. 10 is a block diagram showing an example of the apparatus construction when an image binarizing apparatus of the fifth embodiment of the present invention is used in a digital camera.

13 Fig. 10 is a block diagram showing an example of the apparatus construction when an image binarizing apparatus of the sixth embodiment of the present invention is used in a digital camera.

14 Fig. 10 is a block diagram showing an example of the image capturing apparatus of the seventh embodiment of the present invention when used in a digital camera.

15 Fig. 10 is a flowchart showing the flow of image processing as far as the binarization of a multi-valued image in a digital camera of the seventh embodiment is concerned.

16 Fig. 10 is a block diagram showing an example of the apparatus construction in terms of binarization and recording of an incoming image in a digital camera using a CMOS sensor in the image input section.

17 Fig. 10 is a block diagram showing an example of the image capturing apparatus of the ninth embodiment of the present invention when used in a digital camera.

18A and 18B show relationships between an interpolation block and a screen in a digital camera of the ninth embodiment.

19 Fig. 10 is a flowchart showing the flow of image data processing in terms of binarizing a multi-valued image in a digital camera of the ninth embodiment.

20 Fig. 10 is a structural diagram showing an example when an image capturing apparatus for calculating binarization thresholds applied to each pixel in a predetermined block and performing binarization of the image data for each pixel using a CMOS sensor and a photometer used in a digital camera.

21 Fig. 10 is a block diagram showing an example of the apparatus construction when an image sensing apparatus of the eleventh embodiment of the present invention is used in a digital camera.

22 Fig. 10 is a flow chart showing the flow of image data processing as far as the binarization of a multi-valued image in a digital camera of the eleventh embodiment is concerned.

23 Fig. 10 is a block diagram showing another example of the apparatus construction of a digital camera of the eleventh embodiment.

24 Fig. 10 is a block diagram showing an example of the apparatus construction when an image sensing apparatus of the twelfth embodiment of the present invention is used in a digital camera.

25 Fig. 10 is a flowchart showing the flow of image data processing as far as the binarization of a multi-valued image in a digital camera of the twelfth embodiment is concerned.

26 Fig. 10 is a block diagram showing an example of the apparatus configuration from image input to recording of a binarized image when an image capturing apparatus that performs image binarization in a CPU by software processing is used in a digital camera.

27 Fig. 10 is a block diagram showing the apparatus construction from an input image to a binarized image recording when an image capturing apparatus that performs image binarization on a CPU by software processing is used in a digital camera.

28 Fig. 10 is a diagram showing an example of the construction of a computer system when the present invention is performed using software.

The The present invention will now be described in detail with reference to FIG the drawings described.

The first embodiment, which will be explained below, relates to an image binarizing apparatus of the present invention used in a digital camera. 1 Fig. 10 is a block diagram showing an example of the apparatus configuration from the input of the image data to the recording of a binarized image (binarized data) when an image binarizing apparatus of the present invention is used in a digital camera.

A digital camera 100 includes a CCD 101 , an A / D converter 102 , a white balance adjuster 103 , a pixel interpolator 104 , a luminance generator 105 , an aperture corrector 106 , a frame memory 107 , a CPU 108 , a block buffer 109 , a mean luminance value calculator 120 , a low luminance threshold adjuster 121 , a setting circuit for a binarization threshold 122 , a binizer 123 , a compressor 124 and a frame store 125 ,

The CCD 101 converts light through an optical system, not shown, of the digital camera 100 is collected into electrical signals, and outputs R, G, B analog signals for each pixel forming a multi-valued image as image data. The output analog signals are converted into digital signals by the analog to digital converter 102 converted. The white balance of the digital signals is controlled by the white balance adjuster 103 adjusted or adjusted. In the pixel interpolator 104 For example, the interpolation of R, G or B signals for which there is no information is performed on each pixel with the image data after their white balance is adjusted. Thereafter, R, G, and B are treated as if they are red, green, and blue, or alternatively represent a red signal value, a green signal value, or a blue signal value.

The ratio between the filter of the CCD 101 and the interpolation by the pixel interpolator 104 will now be described. 2 Fig. 10 is a conceptual view of a light receiving portion of the CCD 101 , The R, G and B filters are set in a fixed pattern in the light receiving section, and color differences are identified by these filters. Remarkably, it is normal to provide more filters for green pixels for which the human eye is highly sensitive than for other colors. The suffixes are used as identifiers for identifying a position (filter number). It is also noteworthy that filter numbers are provided only for the central portion of the figure.

A value R (G0) for a red R interpolation signal and a value B (G0) for a blue B interpolation signal at the position G0 are calculated by using the following equation (1): R (G0) = (R0 + R2) / 2 B (G0) = (B0 + B1) / 2 (1)

The values G (R0), B (R0), G (B0) and R (B0) for the green G interpolation signal at the positions R0 and B0 are calculated by using the following equation (2): G (R0) = (G0 + G1 + G2 + G5) / 4 B (R0) = (B0 + B1 + B4 + B5) / 4 G (B0) = (G0 + G1 + G3 + G6) / 4 R (B0) = (R0 + R2 + R4 + R5) / 4 (2)

The pixel interpolator 104 performs the interpolation expressed above, at each pixel position, and outputs the interpolated R, G and B signals for each pixel.

The luminance generator 105 generates a value Y for a luminance signal from each of the interpolated pixels, using the following equation (3): Y = 0.34R + 0.55G + 0.11B (3)

A multiplier and adder are necessary to calculate the luminance signal values, which are digital values, using Equation (3). However, by approximating the equation (3) using the following equation (4), it is possible to use the luminance generator 105 only build up with an adder. Y = (2/8) R + (5/8) G + (1/8) B (4)

Consequently Is it possible, if this equation (4) is used, the values for the luminance signal to calculate the luminance values Y with a simple circuit construction to calculate. Accordingly, it is possible to use a digital camera with low circuit costs and excellent calculation speed and low power consumption.

High pass portions of the image data of the luminance signal value Y are detected by the aperture corrector 106 improved. The aperture corrector is performed by using a conventional 5 × 5 size high pass enhancement filter. After the high-pass enhancement, the signals for the luminance value temporarily become temporal in the frame memory 107 saved.

The CPU 108 calculates the block size and the sampling interval (sampling cycle). The block buffer 109 which is described below, the calculator 120 for the mean luminance value and other circuits and members of the digital camera 100 are also done by the CPU 108 controlled. Because the image size of the multi-valued image is determined by the size of the CCD (ie, by the number of pixels), the CPU calculates 108 the necessary block size and sampling interval (sampling cycle) to binarize the value-added image based on the image data.

An unevenness of the luminance values produced by the optical system of the digital camera 100 are caused (for example, the lens) differ depending on the position and the intensity of the light source, but as a rule, the central area of an image is bright, while the image tends to become darker the farther it is approaching his edge. Accordingly, the CPU calculates 108 the image partitioning pattern taking into consideration peripheral light reduction of the optical system. It is noteworthy that a subdivision pattern is selected from a solid block subdivision pattern set in a not-shown memory section or the like.

3A and 3B show examples of subdivisions of multivalued images into blocks. In addition to a normal subdivision into rectangular shapes shows 3A an example where a multi-valued image is divided into a combination of squares, rectangles, and triangles point-symmetrically about the center of the image (the lens center). In contrast, shows 3B an example where a multi-valued image is divided on the basis of concentrated circles around the center of the image. By dividing the image into blocks taking due account of the optical system, the brightness of each block can be made uniform. In addition, since binarization thresholds are set for block units as described below, high image quality binarization of a multi-valued image becomes possible. Notably, in the description given below, the image is divided into square blocks to simplify the description.

The block buffer 109 reads from the frame memory 107 , which may also be referred to as a frame memory or a frame memory, forms an image in the block units of the block partition pattern generated by the CPU 108 is determined, and temporarily stores the image. It is noteworthy that blocks that are in the block buffer 109 are called object blocks (ie these blocks are the object of subsequent processing).

The calculator 120 for the average luminance value, pixels of images captured in the block buffer 109 are stored, according to a predetermined sampling cycle, and calculate the average luminance value. 4A and 4B show examples of sampling intervals at which the pixels in the blocks are scanned the. 4A shows a state when the size of the image is 1280 pixels by 960 pixels, and the CPU 108 sets the block size at 64 pixels by 64 pixels and the sampling cycle at 2 ( 4A shows only a group of 9 pixels by 8 pixels of a block). In contrast, the shows 4B a state when the size of the image is 2560 pixels by 1920 pixels, and the CPU 108 sets the block size to 128 pixels by 128 pixels and the sampling cycle to 4 ( 4B shows a group of 9 pixels by 8 pixels of a block).

In accordance with the processing performance, while the power consumption is considered, it is also for the CPU 108 possible to set as fixed a total number of blocks for the size of the image to be binarized or a sampling interval within a block. Accordingly, even if the size of the image is larger (that is, even if the total number of pixels is larger), the sample number can be fixed, and the processing time until the binarization threshold is determined can be shortened. As a result, low power consumption binarization processing is possible. It is noteworthy that the CPU 108 set the sampling interval with units of a block.

5 is a block diagram showing an example of the structure of the calculator 120 for a mean luminance value. The luminance value (set here as v) of the pixels passing through the CPU 108 is scanned with a low luminance threshold (here set at th1 (i, j) (th1 (i, j is an index representing the block number, and this block is set at B (i, j))))) through a comparator or comparator 501 compared. If the luminance value v of the sampled pixels is greater than the low luminance threshold thl (i, j), the comparator outputs 501 1, and when the luminance value v of the sampled pixels is lower than the low luminance thresholds thl (i, j), the comparator outputs 501 a signal value of 0.

If a signal value of 1 is output, a gate is opened and the luminance value v becomes the adder 506 entered. The adder 506 adds a value of an addition result register 502 (hereinafter referred to as sumv) to the input luminance value v of the pixels, and the addition result register 502 saves the new addition result. The signal value of 1 from the comparator 501 also becomes a counter 503 which counts the number of luminance values v (referred to hereinafter as num) passing through the gate 507 must arrive.

The above processing can be expressed as an algorithm in a mathematical equation by the following equation (5). if v> thl (i, j) then sumv = sumv + v num = num + 1 further sumv = sumv num = num (5)

If, as a result of the increase, the counter 503 is placed in a state where the number of digits is constantly increasing (ie, in a state where the numerator is expressed as a power of two), the gate becomes 508 opened, and the sum of the luminance values v (sumv) contained in the addition result register 502 is held, becomes a shift register 504 transfer. A bit position with a 1 in the counter 503 begins, is shifted to the right by a circumference of -1. After the processing for all the pixels in a block has been completed, the value stored in the shift register becomes 504 is stored as a mean luminance value ave (i, j). Namely, the average luminance value is calculated by the following equation (6): ave (i, j) = sumv '/ num' (6)

Here, num 'is a numerical value less than the sample number sampled in the block B (i, j) and represents the largest value of a number represented as power of two, while sumv' den Represents value in the addition result register 502 is held when num 'is counted.

The adjuster 121 for the low luminance threshold multiplies a predetermined coefficient by the mean luminance value ave (i-1, j) of the previous neighboring block (ie if the current block is B (i, j) then the previous adjacent block B is i (i-j)). 1, j)) and calculates the low luminance value thl (i, j) obtained in the calculator 120 is to be used for the mean luminance value. If the predetermined coefficient is taken as Ca = 1/4 (1 / power of two), then the adjuster 121 for the low luminance value, a value from which the low order 2 bits of ave (i, j) are removed, and thus no special circuit is required and the circuit construction can be simplified. In addition, high-speed processing with low power consumption becomes possible. The calculation described above can be represented by the following equation (7): thl (i, j) = ave (i-1, j) x Ca (7)

In In this case, the mean luminance value ave became only one adjacent one Block used when the low luminance threshold thl however, it depends on the type of block division just as possible to use the mean luminance value ave of all adjacent blocks (for example those blocks above, below, on the left and on the right).

The adjustment circuit 122 for the binarization threshold uses the mean luminance value ave (i, j) determined by the calculator 120 is calculated for the mean luminance value to set a binarization threshold TH (i, j) to be used for binarizing a multi-valued image. The binarization threshold TH (i, j) is set by setting the average luminance value ave (i, j) by, for example, a predetermined coefficient Cb. In the same way as in the setting circuit 121 for the low luminance threshold, if Cb is set as x / 16 or Cb is set as x / 8 (where x is a predefined value representing a natural number not larger than the denominator), then the setting circuit can be set 122 for the binarization threshold by using only one adder. Accordingly, the circuit construction can be simplified, and high-speed processing with low power consumption becomes possible. If the above is represented as a mathematical equation, the following equation (8) is obtained: TH (i, j) = ave (i, j) × Cb (8)

The binarizer 123 compares each pixel of the block buffer 109 with the binarization threshold TH and performs a binarization. The binarization image is subjected to image compression suitable for a binary image, such as MH or MR ge, in the compressor 124 , The compressed image is then in the image memory 125 saved.

In the above description, the CPU corresponds 108 the block dividing unit and the sampling unit; the CCD 101 , the A / D converter 102 , the white balance adjuster 103 , the pixel interpolator 104 , the luminance generator 105 and the aperture correction 106 correspond to the output unit for the luminance value; the adjustment circuit 122 for the binarization threshold, the binarization threshold setting unit corresponds; the binarizer 123 corresponds to the binarization unit; the adjuster 121 for the low luminance threshold, the low luminance threshold setting unit; the frame memory 107 , the block buffer 109 and the CPU 108 correspond to the object block selection unit; the comparator 501 corresponds to the distance unit for the low luminance value, and the calculator 120 for the average luminance value, the average luminance value calculation unit.

Next, the procedure of processing in terms of the binarization of a multi-valued image will be described. 6 Fig. 10 is a flowchart showing the flow of image data processing until a multi-valued image is binarized. The CPU 108 reads the size of a multi-valued image stored in the frame memory 107 is stored (step S601). It is noteworthy that it is also possible to get information from a pre-set number of pixels in the digital camera 100 or from the CCD 101 or the like, instead of the frame memory 107 , Next, the CPU sets in accordance with the image size 108 the sampling cycle (step S602).

On the basis of the image size read in step S601 and the sampling cycle set in step S602, the CPU sets 108 the block size, shape and subdivision pattern are determined (step S603) because the number of pixels provided by the CCD 101 output, is fixed, or is an estimated pixel number, such as 640 × 480 pixels or 800 × 600 pixels specified by a setting mode, it is also possible for the CPU 108 selects a predetermined sampling cycle and a predetermined block shape.

Next, a single block B (i, j) of the plurality of generated blocks is designated, and the image information becomes the block buffer 109 transferred (step S604). The pixels (Nominanzwerte) are then sampled by the block B (i, j) in accordance with the sampling cycle, which is determined in step S602 (step S605). Next, on the basis of the mean luminance value ave (i-1, j) of the adjacent block B (i-1, j) calculated in the previous routine, a low luminance threshold value thl (i, j) for the Blocks B (i, j) are calculated, and luminance values that do not reach the low luminance threshold thl (i, j) are then removed from the luminance values of the sampled pixels (step S606). The removal of the low luminance pixels in step S606 assists in high quality binarization processing.

Using the luminance values from which the low luminance values have been removed, the average luminance value ave (i, j) of the block B (i, j) is calculated (step S607). The binarization threshold TH (i, j) for block B (i, j) becomes then calculated on the basis of the mean luminance value ave (i, j) (step S608). The binarization threshold TH (i, j) is then used to binarize all the pixels (used as g (x, y)) of the block B (i, j) stored in the block buffer 109 is stored (step S609). Notably, x and y are natural numbers representing the position of each pixel within the block.

When Lastly, a determination is made whether or not the binarization processing is performed for all blocks has been completed (step S610). When all blocks are binarized (i.e., the result of the determination in step S610 YES), then the processing ends. If not all blocks have been binarized are (that is, the result of the determining step in step S610 is NO), a block adjacent to the block B (i, j) becomes is (for example, the block B (i + 1, j)) fixed; and steps S604 to 5610 are repeated.

at the first embodiment An example has been described when the image binarizing apparatus of the present invention in a digital camera has been. The block division is performed to match the image size and the optical system and samples of the luminance values be from the blocks ext rahiert. Because of that and because of a low threshold for luminance binarization under consideration of surrounding blocks is fixed and the binarization based on it It is, even if all the pixels within a block part a big one Character or character or if the brightness is between blocks changes abruptly, possible, calculate a suitable mean luminance value and a fast mean luminance value Binarization of a High quality picture perform at low power consumption. This in turn allows that the usable duration of the battery of the digital camera be extended can.

The second embodiment explained below relates to an image binarizing apparatus of the present invention used in a digital camera. The 7 Fig. 10 is a block diagram showing an example of the apparatus construction from the image input until a binarized image is recorded in a digital camera using a CMOS sensor in the image input section. It is noteworthy that in the present embodiment, those structural portions that are identical to those of the first embodiment are given the same symbols and a detailed description is therefore omitted. Here, those portions which are different from those of the first embodiment will be mainly described.

The digital camera 700 comes with a CMOS sensor 701 provided in the image input section. As a result, the CMOS sensor is 701 unlike the CCD 101 (please refer 1 ) capable of only one raster scan capable of random access, and capable of reading in block units. That's why the frame store is 107 not required, and the circuit structure can be simplified. Unlike the CCD 101 which requires a separate power source for the CMOS integrated circuits, which are the other circuits of the digital camera 700 form, the CMOS uses 701 the same power source as the CMOS integrated circuits, reducing power consumption. Because the size of the circuit that the digital camera 700 As a result, it is also more practical than a system using a CCD in terms of power consumption, processing speed, and cost. Notably, a CMOS sensor is used in the present embodiment, however, another image binarizing apparatus having one through a block accessible image input portion may also be used.

In this third embodiment, a description will be made as to when an image binarizing apparatus is used in a digital camera which calculates binarization thresholds to be applied to each pixel at a predetermined block and calculates the image data for each pixel. 8th Fig. 10 is a block diagram showing an example of the apparatus of the digital camera which calculates binarization thresholds to be applied to each pixel and calculates the image data for each pixel. Notably, in the present embodiment, those structural portions identical to those in the first embodiment are denoted by the same reference numerals and symbols, respectively, and a detailed description thereof will be omitted. Here, those portions which are different from those of the first embodiment will be mainly described.

The digital camera 800 comes with a setting circuit 801 for a block binarization threshold into which average luminance values are input by the calculator 120 for the average luminance value, and outputs the binarization thresholds to be applied for each block (these will be referred to as block binarization thresholds hereinafter); a memory 802 for storing the block binarization thresholds; and a binarization threshold interpolator 803 which sets binarization thresholds to be applied for each individual pixel in a predetermined block based on the block binarization thresholds. Apart from that, the structure is the same as that of the digital camera 100 the first embodiment. It is noteworthy that a section for generating color difference signals is omitted.

The adjustment circuit 801 for the block binarization threshold is with the setting circuit 122 for the binarization threshold of the digital camera 100 identical (see 1 ) and gives block binarization thresholds for object blocks for storage in memory 802 out. The adjustment circuit 801 for the block binarization threshold calculates block binarization thresholds for all blocks that are sequential in memory 802 to save.

The interpolator 803 for the binarization threshold uses the block binarization thresholds of all blocks that are in memory 802 are stored to set the binarization thresholds to be applied to each pixel at a predetermined range. Notably, in the description given below, the term interpolation block refers to a predetermined range, with a binarization threshold being set for each pixel forming that pixel. A description will now be given on the nature of the calculation of the binarization thresholds to be applied to each pixel in the interpolation block. 9A and 9B Fig. 10 are explanatory diagrams summarizing the calculation of binarization thresholds to be applied to each pixel within one interpolation block. 9A shows the relationship between object blocks and an interpolation block, and 9B Figure 12 shows the calculation of the binarization thresholds to be applied for each pixel within an interpolation block.

How clear in 9A is shown, the interpolation block BH bridges four adjacent object blocks Ba, Bb, Bc and Bd. The binarization thresholds of the object blocks Ba, Bb, Bc and Bd are respectively set as a, b, c and d. The binarization threshold interpolator 803 connects the block binarization thresholds a, b, c and d to calculate the binarization thresholds to be applied for each pixel within the interpolation block BH.

The following is a description of the method of calculating binarization thresholds to be applied to each pixel bp within the interpolation block BH, wherein FIG 9B Reference is made. It is assumed that the shape of the interpolation block BH is rectangular, with the size (the number of pixels) thereof being set as xbnum in the horizontal direction and ybnum in the vertical direction. The position of the pixel bp is set as (m, 1). At this time, it may be assumed that the value a is at the point (0, 0) of the interpolation block BH, the value b at the point (xbnum 0) of the interpolation block BH, the value c at the point (0, ybnum) of the interpolation block is BH and the value d is at the point (xbnum, ybnum) of the interpolation block BH. The area between these points is approximated linearly.

If the temporary threshold on the boundary line on the left side of the interpolation block BH is taken as lefth and the temporary threshold is taken on the boundary line on the right side of the interpolation block BH as rightth, then the following equation (9) can be applied first: leftth = (a (ybnum - 1) + c1) / ybnum rightth = (b (ybnum - 1) + d1) / ybnum (9)

Next, assuming that the point (0, 1) has the value leftth and the point (xbnum, 1) has the value rightth, then the binarization threshold th (m, 1) becomes the pixel bp is linearly approximated using the following equation (10): th (m, 1) = (leftth (xbnum - m) + rightth × m) / xbnum (10)

Remarkable is that, if the interpolation block BH at the edge or There is only one or two adjacent to the edge of the entire image object blocks gives. In this case, you can by substituting the obtained block binarization thresholds for the unavailable thresholds for the Block binarization equations (9) and (10) are used. For example at the upper left edge or upper left edge of the image the only available threshold for the Block binarization d and the thresholds a, b, and c for block binarization unavailable. As a result, in this case, the value d is substituted for a, b and c.

In the binarizer 123 the luminance values of the pixels bp in the interpolation block obtained from the frame memory 107 or the block buffer 109 is obtained, compared with the binarization thresholds for the pixels bp passing through the interpolator 803 is calculated for the binarization thresholds, and the binarization of the luminance values of the interpolation block BH is performed. An image compression performed on a binarized image, such as MH or MMR, is performed on the binarized image data by the compressor 124 in the same way as used in the first embodiment.

Next, a description will be given of the process flow until a multi-valued image has been binarized in the present embodiment. 10 Fig. 10 is a flowchart showing the flow of image data processing until the multi-valued image has been binarized. The CPU 108 reads the size of the multivalue image that is in frame buffer 107 has been stored (step S1001). It is noteworthy that it is possible to get the information from the CCD 101 to use or a pixel number in the digital camera 100 was preset, or the like, instead of the frame memory 107 , Next, the CPU sets in accordance with the image size 108 the sampling cycle (step S1002).

On the basis of the image size read in step S1001 and the sampling cycle set in step S1002, the CPU sets 108 the block size, shape and subdivision pattern are determined (step S1003). Because the number of pixels used by the CCD 101 It is the same for the CPU to output, set, or be a pixel number such as 640 × 480 pixels or 800 × 600 pixels specified by a switching mode 108 possible to select a predetermined sampling cycle or block shape.

Next, a single block B (i, j) is set by the plurality of blocks that have been generated, and image information becomes the block buffer 109 transferred (step S1004). Pixels (luminance values) are then sampled by the block B (i, j) in accordance with the sampling cycle set in step S1002 (step S1005). Next, on the basis of the mean luminance value ave (i-1, j) of the adjacent block B (i-1, j) calculated in the previous routine, the luminance threshold value thl (i, j) for the block is calculated B (i, j) is calculated. The luminance values that do not reach the low luminance threshold th1 (i, j) are then removed from the luminance values of the sampled pixels (step S1006). The removal of the low luminance pixels in step S1006 helps to achieve a high quality binarization process.

Using the luminance values from which the low luminance values have been removed, the average luminance value ave (i, j) of the block B (i, j) is calculated (step S1007). The block binarization threshold value TH (i, j) for the block B (i, j) is then calculated based on the average luminance value ave (i, j) (step S1008). The calculated block binarization threshold TH is then stored in memory 802 stored (step S1009). A determination is then made as to whether or not the thresholds for block binarization have been calculated for all the blocks (step S1010). If the calculations have not been completed (ie, if the result of the determination in step S1010 is NO), a block adjacent to block B (i, j) (for example, block B (i + 1, j)) is set, and steps S1004 to S1010 are repeated.

If thresholds for the block binarization for all blocks have been calculated (the result of the determination in step S1010 is YES), an interpolation block is set (step S1011). This interpolation block is capable of performing a high quality sub binarization or the like and may be specified in advance by the user, or may be determined by an appropriate mode switching so as to cover the central portion of the image. The interpolator 803 for the binarization threshold uses the threshold TH for the block binarization of the object blocks that are bypassed by the interpolation block set in step S1011 to set the binarization thresholds th (x, y) of each individual pixel within the interpolation block (Step S1012). It is noteworthy that coefficient f in 10 conceptually represent equations (9) and (10) while x and y are natural numbers representing the positions of each pixel in an image.

The pixels (assumed here to be g (x, y)), which are from the frame memory 107 are read, are binarized by the binarizer 123 using the binarization thresholds calculated in step S1011 (step S1013). Next, a determination is made as to whether or not binarization processing has been performed on all the pixels (S1014). If binarization processing has not been performed on all pixels (ie, if the result of the determination in step S1014 is NO), then the pixel adjacent to the pixel g (x, y) (for example, the pixels g (x + 1, y )), and steps S1012 to S1014 are repeated. If the determination is made that the binarization process has been completed for all the pixels (ie, if the result of the determination in step S1014 is YES), a determination is made as to whether or not the binarization process or the binarization processing for all interpolation blocks have been completed (step S1015). If the binarization processing has not been completed (ie, if the result of the determination in step S1015 is NO), then steps S1011 to S1015 are repeated. When the binarization processing has been completed (If the result of the determination in step S1015 is YES), then the process ends.

at the third embodiment the block partitioning is performed in such a way that it is the Bit size and the optical system, and samples of the Luminance values are given by the blocks extracted. Because of that and because of a low threshold for luminance binarization considering the surrounding blocks was set and the binarization based on it will, and in addition because the binarization carried out is set by setting binarization thresholds which apply to every pixel in an interpolation block with the above block binarization thresholds used, high-quality image processing becomes possible by the digital camera of the first embodiment is used. In particular, if a notebook or a note or the like It is when all the characters are small or if it is photographed Characters in a section of the area or area are small, possible, a partial binarization to perform in those places of high quality.

The third embodiment discussed below relates to an image binarizing apparatus of the present invention which uses a CMOS sensor when calculating binarization thresholds to be applied to each pixel of a predetermined block, and to binarizing the image data for each pixel in a digital camera performs. 11 Fig. 10 is a block diagram showing an example of the apparatus of a digital camera using a CMOS sensor in the image input section, calculating binarization thresholds to be applied to the pixel, and binarizing the image data for each pixel. Notably, in the present embodiment, those structural portions that are identical to those in the third embodiment have the same reference numerals, and a detailed description thereof will be omitted. Mainly those portions that are different from those in the third embodiment will be described.

The digital camera 1100 is with a CMOS sensor 1101 at in the image input section. As a result, it's different than the CCD 101 (please refer 8th ), which can only perform one raster scan, the CMOS sensor 1101 thereby being able to randomly access, read a block as a single unit, read pixel values for each block, calculate block binarization thresholds, and data in memory 802 save. Therefore, the block buffer becomes 109 (please refer 8th ) is not needed, and the circuit construction can be simplified.

At the same time, the pixel values read per block are converted to luminance signals, subjected to aperture correction, and are stored in the frame memory 107 saved. Next used differently than the CCD 101 which requires a separate power source for the CMOS integrated circuits, which are the other circuits of the digital camera 800 form, uses the CMOS sensor 1101 the same power source as the integrated CMOS circuits, thereby reducing power consumption. Because the size of the circuit that the digital camera 1100 As well as being reduced, it is more practical than a system using a CCD in terms of power consumption, processing speed and cost. It is noteworthy that a CMOS sensor is used in the present embodiment, however, another image binarizing apparatus having an image input section that can be accessed on a block-by-block basis may also be used.

In the fifth embodiment, a description will be given of the case where an image binarizing apparatus is used in a digital camera in which the spread of the average luminance values is restricted so that the values are kept within a predetermined range. 12 Fig. 10 is a block diagram showing an example of the apparatus of a digital camera in which the spread of the average luminance values is limited so that the values are kept within a predetermined range. It is noteworthy that those structural portions of the present invention which are identical to those of the first embodiment have been given the same reference numerals and symbols, and the description thereof is omitted.

The digital camera 1200 is with a limiter 1201 which limits the average luminance values provided by the calculator 120 for the average luminance values using a preset low limit value so as to prevent the average luminance values, which are too low, from being output. As an example, relatively bright characters may be used in a dark image, a more specific example may be the image of a cat drawn with white chalk on a (black) board. Compared to the surface of the board, the area of the white chalk figure is extremely small, but in this case the important information of the picture is the cat drawn with white chalk. If normal binarization is performed in this case, the image data of the portions of the panel may become dominant except those portions which contain the cat's image in white chalk. As a result, will last finally set the binarization threshold at the boundaries between light and dark of the black sections of the panel. Consequently, the white chalk image is naturally judged white, but chalk dust and chalk pieces (for example, a trace left by a table sponge or the like) are also judged white, and the end result is that the noise increases.

The limiter 1201 sets average average luminance values so that this type of noise is not reproduced. Specifically, a determination is made as to whether or not there are any average luminance values within a predetermined range, and if the result of this determination is negative, then the mean value for the low limit or the upper limit of that range is substituted. By setting the lower limit (or upper limit) of the mean luminance value, the limiter 1201 is used, it is possible to obtain a binarized image with high contrast. Depending on the mode of use, it is also possible to limit the average luminance value by a predetermined upper limit, so that an average luminance value that is too high is not output.

Notably, in the present embodiment, a CCD 101 however, depending on the mode of use, it is also possible to use a CMOS sensor 701 to use (see 7 ), as is the case with the second embodiment.

In this sixth embodiment, the case where a digital camera uses an image binarizing apparatus in which binarization thresholds applied to each pixel at a predetermined one are calculated and binarization of the image data is performed for each pixel will be described Spread of average luminance values is limited to be within a predetermined range. 13 Fig. 10 is a block diagram showing an example of the construction of a digital camera according to the present embodiment. Notably, portions of the structure of the present embodiment that are identical to those of the fifth embodiment have been given the same symbols and a description thereof will be omitted.

The digital camera 1300 is with a limiter 1301 which limits the average luminance values provided by the calculator 120 are outputted for the mean luminance values using a preset upper limit value so as to prevent average luminance values that are too low from being output. The mean luminance values given by the limiter 1301 are output within a predetermined range. For example, average luminance values that protrude due to being affected by a reflected flash from a panel or the like are set to fall within a predetermined range. Accordingly, block binarization thresholds determined by the setting circuit can be prevented 801 are set for block binarization thresholds, do not match or are inconsistent with the values of the other blocks.

More specifically, in the present embodiment, block binarization thresholds to be applied to each pixel in one interpolation block are determined by the interpolator 803 for block binarization thresholds by using the block binarization thresholds of object blocks in the environment. Because average luminance values that protrude or greatly differ through the limiter 1301 can be removed, each block binarization threshold can be set more accurately at that time. As a result, even higher quality binarization processing becomes possible. Depending on the mode of use, it is also possible for the average luminance values to be limited by a lower limit set in advance so that average luminance values that are too low are not output. In addition, the limiter 1301 also set a predetermined range by using the block binarization thresholds.

Notably, in the present embodiment, a CCD 101 however, depending on the mode of use, it is also possible to use a CMOS sensor 1001 to use (see 11 ), as in the fourth embodiment.

The seventh embodiment, which will be explained below, refers to an image binarizing apparatus having a photometric unit. 14 Fig. 10 is a block diagram showing an example of the apparatus construction from image input to recording of a binarized image when an image sensing apparatus equipped with a photometer is used in a digital camera. Notably, in the present embodiment, structural members similar to those of the first embodiment are given the same reference numerals and symbols, respectively, because the structural members are similar to those of the first embodiment, and a detailed description of these members is omitted. Those portions that are different from the first embodiment will be mainly described.

The digital camera 1400 includes a CCD 101 , the A / D converter 102 , the white balance adjuster 103 , the pixel interpolator 104 , the luminance generator 105 , the aperture corrector 106 , the frame memory 107 , the CPU 108 , a photometer 1401 , a smoother 1402 , a store 1403 , a block read control device 1404 , a circuit 122 for setting a binarization threshold, the binarizer 123 , the compressor 124 and the image memory 125 ,

The CCD 101 converts light through an optical system, not shown, of the digital camera 1400 is collected into electrical signals, and outputs R, G, B analog signals for each pixel, forming a multi-valued image. The output analog signals are converted into digital signals by the A / D converter 102 converted. If you go through the white balance adjuster 103 , Pixel interpolator 104 , the luminance generator 105 and the aperture corrector 106 If so, the digital signals are subjected to processing such as luminance value interpolation, extraction, and the like, and are then stored in the frame memory 107 temporarily saved.

Based on the luminance information from the photometer 1401 (described below) controls the CPU 108 the block read control device 1404 (described below) and divides an image stored in the frame memory 107 is stored, in the manner that is the same as the screen division made by the photometer 1401 used for photometry. The screen partition used for the photometry from the photometer 1401 can be used as well. Alternatively, it can also be done by controlling the CPU 108 be done in the same way as in 3 , The CPU 108 controls the other circuits and links of the digital camera 1400 ,

The photometer 1401 is equipped with an automatic exposure detection mechanism to perform photometry on an object to be photographed before photographing an image, and measures the brightness of each screen based on digital signals received from the A / D converter 102 be issued. The method of photometry may include, for example, a measurement in which the luminance values of the pixels are added. At this time is the same for the CPU 108 possible, photometric values for the photometer 1401 to key or to be used, which are to be used in the addition. It is noteworthy that, depending on the mode of use, it is also possible to use a quantity of light that is photoelectrically transmitted through the CCD 101 was converted to use.

The smoothing device 1402 or the smoother 1402 smoothes the photometric values of each screen or measuring range taken by the photometer 1401 and gives it as ave (i, j) to the memory 1403 out. The processing shown below is an example of this smoothing. If the average value ave (i, j) of the photometric values of all the pixels in a screen (this screen is referred to herein as G (i, j)) within the photometer 1401 Alternatively, the average value of the photometric values scanned within the screen G (i, j) is inconsistent with the mean values of the photometric values of the screens in the environment, the photometric values of each pixel of the screen become G (i, j) is corrected so as to be consistent with the average values of the photometric values of the screens in the environment.

An example of an algorithm for performing this processing is given below. The average photometric value of the four screens adjacent to the screen G (i, j) is set as ave4 (i, j). If the average photometric value ave (i, j) of the screen G (i, j) is three times the value ave4 (i, j) or more, then the photometric values of each pixel in the screen G (i, j) ( which is referred to herein as s (x, y)) is converted by using the following equation (11): if ave (i, j) ≥ 3 × ave4 (i, j) then s (i, j) = ave4 (i, j) + (1/4) x (s (i, j) - ave4 (i, j)) (11)

When the smoother 1402 For example, the converted photometric values are used to recalculate the average photometric value ave (i, j) of the screen G (i, j), and the mean photometric value then becomes the memory 1403 output. The memory 1403 stores this mean photometric value.

The block read control device 1404 subdivides the images in the frame memory 107 are stored, in the same manner as the screen subdivision, for the photometry of the photometer 1401 is used. As a result, the binarization thresholds are made more natural.

The adjustment circuit 122 for binarization thresholds sets the binarization value TH (i, j) based on ave (i, j). The adjustment unit represents a multiplier for multiplying ave (i, j) by a preset coefficient Cb, however, if the coefficient Cb is made equal to x / 16 or Cb is made equal to x / 8 (where x is a predetermined value which is a natural value) Number not larger than the denominator), then the binarization threshold setting circuit can be easily constructed from an adder that is cheap in both cost and speed.

Based on the binarization threshold TH (i, j), the binarizer binaries 123 each pixel of the block B (i, j), corresponding to the screen G (i, j). Remarkably, smoothed blocks are binarized by appropriate processing using converted photometric values. The compressor 124 performs compression suitable for a binary image, such as MH or MR. The compressed image is then stored in the image memory 125 saved.

In the above description, the CCD corresponds 101 the image capture unit; the CPU 108 and the photometer 1401 correspond to the screen dividing unit; the photometer 1401 corresponds to the photometry unit; the CPU 108 corresponds to the block dividing unit; the circuit 122 for setting the binarization threshold value corresponds to the binarization threshold setting unit; the binarizer 123 corresponds to the binarization unit; and the smoother 1402 corresponds to the smoothing unit for the photometric value.

Next, the flow of image data processing in terms of the binarization of a multi-valued image will be described. 15 Fig. 10 is a flow chart showing the flow of data processing as far as the binarization of a multi-valued image is concerned. First the CPU reads 108 the size of the image from the CCD 101 (Step S1501). Next is the CPU 108 a screen G (i, j) from the photometric values obtained from the photometer 1401 are output (step S1502) and calculate the average value of the photometric values of that screen (step S1503). The average can be determined by using all photometric values or by using an appropriate sample thereof.

Next, a determination is made as to whether or not the average value of the photometric values of the screen is inconsistent with the averages of the photometric values of the adjacent screens (step S1504). If the mean value is inconsistent (ie, the result of the determination in step S1504 is YES), then the mean value of the photometric values is smoothed by the smoother 1402 is used (step S1505). If the mean value is not inconsistent (ie, the result of the determination in step S1504 is NO), or alternatively, if the smoothing processing of step S1504 has been completed, the block read control unit reads 1404 the block B (i, j) corresponding to the screen G (i, j) from the frame memory 107 (Step S1506) calculates the binarization threshold based on the photometric value of the screen G (i, j) (Step S1507). The multivalued image within that block is then binarized by using this binarization threshold (step S1508).

A Determination is made whether or not one binarization all blocks has been completed (step S1509). If all blocks are binarized (i.e., if the result of the determination in step S1509 is YES), the binarization processing becomes completed. If not all blocks binarizes (i.e., if the result of the determination in step S1509 is NO), the routine returns to step S1502 which next Screen (for example, the screen G (i + 1, j)) is set and the subsequent steps S1502 to S1509 are repeated.

at this seventh embodiment exists as a binarization threshold is adjusted by using information (photometric values) which are obtained from an automatic photometry section No provided in a digital camera desire after a separate processing to set the binarization threshold. As a result, will simplifies the circuit structure and reduces costs. At the same time, the power consumption can be reduced as a separate processing for calculating binarization thresholds is not necessary. additionally is a binarization high quality the multivalued image possible, even if the image has point reflections from a light source.

The eighth embodiment explained below relates to an image sensing apparatus using a CMOS sensor. The 16 Fig. 10 is a block diagram showing an example of the apparatus structure until an input image is binarized and recorded in the image input section thereof by a digital camera using a CMOS sensor. Notably, in the present embodiment, those portions which are identical to the portions of the seventh embodiment are denoted by the same reference numerals and symbols as in the seventh embodiment and the description thereof is omitted. Herein, those portions which are different from those of the seventh embodiment will be described. The digital camera 1600 is with a CMOS sensor 1601 equipped in the image input section of the camera. Accordingly, unlike the CCD 101 that can only perform a raster scan (s. 14 ), the CMOS sensor 1601 being able to randomly access and read individual block units. That's why the frame store 107 (Frame memory) and the block read control device 1404 unnecessary, which makes it possible to simplify the circuit design. In this case, the CPU performs 108 the function of the block read control device by.

In contrast to the CCD 101 , which requires a separate power source or power source for the CMOS integrated circuit, is the CMOS sensor 1601 able to use the same power source as the integrated CMOS circuit. This allows the power consumption to be reduced. As a result, the power consumption of this system is lower, the processing speed is higher, and the cost is lower than a system using a CCD, since the size of the scale of the digital camera circuit 1600 can also be reduced. It is noteworthy that a CMOS sensor was used in the present embodiment. However, another image capturing apparatus having an input section for blockwise image access may be used.

The ninth embodiment explained below relates to an image capturing apparatus equipped with a photometric unit which calculates binarization thresholds to be applied to each pixel in a predetermined block and performs binarization of image data for each pixel. wherein the image capture apparatus is used in a digital camera. 17 Fig. 13 is a view showing an example of the structure when an image capturing apparatus equipped with a photometer and setting a binarization threshold for each pixel is used in a digital camera. Notably, those structural elements that are identical to those of the seventh embodiment are given the same reference numerals and symbols, and description thereof is omitted because the structural elements of the present embodiment are similar to those of the seventh embodiment. The description given here concerns those parts which are different from the seventh embodiment.

The digital camera 1700 is with a circuit 1701 for setting screen binarization thresholds to which average values of photometric values are inputted from the smoother 1402 are issued and stored in the memory 1403 and which serves to output values multiplied by a predetermined coefficient (hereinafter referred to as screen binarization thresholds) for each screen; the camera is on with a memory 1702 equipped to store screen binarization thresholds; and a binarization threshold interpolator 1703 to set binarization thresholds to be applied to each individual pixel at a predetermined block, based on the screen binarization thresholds. It is noteworthy that a section for generating color difference signals is omitted.

The adjustment circuit 1701 for screen binarization thresholds multiplies a coefficient corresponding to a particular screen with a smoothed photometric value to simplify the following description, the screen having the smoothed photometric value a is referred to as Ga, while the predetermined coefficient is designated Cb (FIG. Ga) is called. The adjustment circuit 1701 for the screen binarization threshold calculates the screen binarization threshold for all screens and stores them sequentially in the memory 1702 ,

The binarization threshold interpolator 1703 uses the screen binarization thresholds for all screens in the memory 1702 are stored to set binarization thresholds to be applied to each pixel in the interpolation block BH. In the description given below, binarization thresholds to be applied to each pixel are set. It is noteworthy that the interpolation block BH by the block read control device 1404 however, depending on the mode of use, it is also possible for the interpolation block BH to be set by the CPU 108 is set or fixed.

The calculation of the binary threshold to be applied to each pixel in an interpolation block will now be roughly described. 18A and 18B show relationships between an interpolation block and screens. It is clearly shown in the figures that the interpolation block BH bridges four neighboring screens Ga, Gb, Gc and Gd. It is noteworthy that the mean values of the smoothed photometric values of the screens Ga, Gb, Gc and Gd are respectively set as a, b, c and d. The interpolator 1703 The binarization threshold uses the smoothed averages of the photometric values a, b, c, and d to calculate the binarization threshold to apply to each pixel within the interpolation block BH.

in the Following is the method of calculating the binarization threshold described on a pixel bp within the interpolation block BH is to apply. The shape of the interpolation block BH becomes assumed to be rectangular, with the size (the number of pixels) of it on xbnum in the horizontal direction and on ybnum in the vertical one Direction is determined. The position of the pixel bp is called (m, 1). At this time, the value a × Cb (Ga) at the point (0, 0) of the interpolation block BH, the value b × Cb (Gb) is at the point (xbnum 0) of the interpolation block BH, the value c × Cb (Gc) is at the point (0, ybnum) of the interpolation block BH, and the value d × Cb (Gd) is at the point (xbnum, ybnum) of the interpolation block BRA. The area between these points is linearly approximated.

at the seventh embodiment the coefficient is given the fixed or fixed value Cb, however, in this ninth embodiment the coefficient differs for each screen. This is necessary a binarization higher Quality too enable, as if the mean of the photometric values of each screen is used. In particular, when the block partition and the Screen partitioning are different, a screen binarization threshold high quality or set.

Of the Binarizing threshold th to be applied to the pixel bp is calculated by the previously mentioned Equation (10) is used. It is noteworthy that in the equations (9) and (10) a, b, c and d by a × Cb (Ga), b × Cb (Gb), c × Cb (Gc) and dxCb (Gd) to be replaced.

At the binarizer 123 are luminance values of the pixels bp in the interpolation block obtained from the frame memory 107 or the block buffer 109 is compared with the binarization threshold for the pixel bp passing through the interpolator 1703 is calculated for the binarization threshold. The binarization of the luminance values of the block BH is then performed. The binarized image data is then subjected to image compression suitable for a binarized image, such as MH or MMR, from the compressor 124 in the same manner as in the seventh embodiment.

19 Fig. 10 is a flowchart showing the flow of image data processing to the point where a multi-valued image is binarized. First the CPU reads 108 the size of the image from the CCD 101 (Step S1901). Next, the CPU sets 108 a screen G (i, j) among the photometric values obtained from the photometer 1401 are output (step S1902), and then calculate the average value of that screen (step S1903). The average can be determined by using all photometric values or by using an appropriate sample thereof.

Next, a determination is made as to whether or not the average value of the photometric values of the screen is inconsistent with the average values of the photometric values of the adjacent screens (step S1904). If the mean value is inconsistent (ie, the result of the determination in step S1904 is YES), then the average value of the photometric values is calculated using the smoother 1402 smoothed (step S1905). If the mean value is not inconsistent (ie, the result of the determination in step S1904 is NO), or alternatively, if the smoothing process has been completed in step S1904, the umbrella binarization threshold is calculated based on the average value of the photometric values (step S1906 ). This screen binarization threshold is then stored in memory 1702 stored (step S1907).

A Determination is then made as to whether or not the screen binarization all blocks has been completed (step S1908). If the screen binarization all blocks completed (i.e., if the result of the determination in step S1909 NO), the routine returns to the step S1902 which next screen (for example, the screen G (i + 1, j)) is then set and the Step S1902 to S1908 are repeated. If the screen binarization all blocks has been completed, (i.e., if the result of the determination in Step S1908 is YES), the interpolation block is set (step S1909). This interpolation block is suitable for partial binarization high quality perform, and can be set in advance by the user, or the central section of the image can be via a suitable mode switching be determined.

The interpolator 1703 for the binarization threshold, the screen binarization thresholds (for example, a × Cb (Ga)) of the screens bridged by the interpolation blocks set in step S1909 uses the binarization thresholds th (x, y) of one of each pixel in the interpolation block (step S1910).

A pixel (referred to herein as g (x, y)) that is out of frame memory 107 through the interpolator 123 is read out is binarized by using the binarization threshold calculated in step S1910 (step S1911). Next, a determination is made as to whether or not the binarization processing has been performed on all the pixels (step S1912). If the binarization has not been performed with all the pixels (if the result of the determination in step S1912 is NO), a pixel in the interpolation block adjacent to g (x, y) is set, for example, g (x + 1, y)), and the steps S1910 to S1912 are repeated. If the binarization has been performed for all the pixels (ie, if the result of the determination in step S1912 is YES), a determination is made as to whether or not binarization for all the interpolation blocks has been completed (step S1913). If the binarization has not been completed for all the interpolation blocks (ie, if the result of the determination in step S1913 is NO), then steps S1909 to S1913 are repeated. If the binarization for all the interpolation blocks has been completed (ie, if the result of the determination in step S1913 is YES), then the processing is ended.

at this ninth embodiment is a separate processing to set a binarization threshold, not necessary because a binarization threshold is adjusted by using information (photometric values) which are obtained from an automatic photometric section which is provided in the digital camera. Consequently For example, the circuitry can be simplified and the cost can be reduced become. Because a separate processing for counting binarization thresholds is not necessary, can simultaneously reduce the power consumption become. Even if the picture is point reflections from a light source has, in addition a binarization high quality of multivalue image possible. Because binarization thresholds, to be applied to each pixel in an interpolation block, be set and a binarization is carried out on the basis of photometric values, Furthermore, image processing of higher quality is possible than with the digital camera of the seventh embodiment. If a sticky note or the like is photographed when all the characters are small or if characters in a section of the area are small, it is particular possible, a partial binarization higher quality to perform at those posts.

The tenth embodiment explained below relates to an image capturing apparatus that calculates binarization thresholds to be applied to each pixel at a predetermined block, and performs binarization of the image data for each pixel, using a CMOS sensor and a photometer used with a digital camera. The 20 13 is a structural diagram showing an example of a case where the image capturing apparatus that computes the binarization thresholds to be applied to each pixel in a predetermined block, and the binarization of the image data for each pixel using a CMOS Sensors and a photometer is used in a digital camera. It is noteworthy that those portions of the present embodiment which are identical to those of the ninth embodiment are denoted by the same symbols and the description thereof is omitted. The description given here mainly concerns those parts which are different from the ninth embodiment.

The digital camera 2000 is with a CMOS sensor 2001 in the image input section thereof. Unlike the CCD 101 (please refer 17 ), which can perform only one raster scan, is because of the CMOS sensor 2001 is able to randomly access and read individual block units, the frame memory 107 and the block read control device 1404 unnecessary. This makes it possible to simplify the circuit construction. In this case, the CPU performs 108 the functions of the block read control by.

Next is in contrast to the CCD 101 that requires a separate power source for the CMOS integrated circuit, the CMOS sensor 2001 capable of using the same power source as the CMOS integrated circuit. This allows a reduction in power consumption. Because the size or dimensions of the circuit of the digital camera 2000 can also be reduced, this system requires less power, the processing speed is higher and the cost is lower than the system using a CCD. It is noteworthy that a CMOS sensor was used in the present embodiment, however, any other image sensing device having an image input section accessible on a block-by-block basis may be used.

The eleventh embodiment explained below relates to an image capturing apparatus in which a binarization threshold is set for each object block based on photometric values output from a photometric unit after those photometric values are adjusted so that they are within a predetermined range. The image capture apparatus is used in a digital camera. The 21 Fig. 10 is a block diagram showing an example of the apparatus construction, from the image input to a binarized image, which is recorded when an image capturing apparatus is used in a digital camera. In the image capturing apparatus, a binarization threshold is set for each object block based on photometric values output from a photometric unit, after those photometric values have been set to be within a predetermined range.

The digital camera 2100 is equipped with the following: the CCD 101 , the A / D converter 102 , the white balance adjuster 103 , the pixel interpolator 104 , the luminance generator 105 , the aperture corrector 106 , the frame store 107 , the CPU 108 , a photometer 2101 , a smoother 2102 , a limiter 2105 , a store 2103 a block read controller 2104 , the binarization threshold setting circuit 122 , the binarizer 123 , the compressor 124 and the image memory 125 , It is noteworthy that a section for generating color difference signals is omitted.

The photometer 2101 is equipped with an automatic exposure detection mechanism (AE detection mechanism) to perform photometry on an object to be photographed before the image is photographed. The photometer 2101 measures the brightness of each screen based on digital signals received from the A / D converter 102 be issued. The method of photometry may, for example, comprise a measurement by adding the luminance values of the pixels. At this time, the CPU can 108 also sample photometric values obtained by adding the photometer 2101 to be used.

The smoother 2102 smoothes the photometric values of each screen from the photometer 2101 and gives it as ave (i, j) to the delimiter 2105 out. The processing discussed below is an example of this smoothing. If the average ave (i, j) of the photometric values of all the pixels in a screen (this screen is referred to herein as G (i, j)) within of the photometer 2101 (alternatively, the average of the photometric values scanned within the screen G (i, j)) is inconsistent with the mean values of the photometric values of the surrounding screens, the photometric values of each pixel of the screen G (i, j ) so corrected that to be consistent with the mean values of the photometric values of the surrounding screens.

An example of an algorithm to perform this method is given below. The average photometric value of the four screens adjacent to the screen G (i, j) is set to ave4 (i, j). If the average photometric value ave (i, j) of the screen G (i, j) is three times the value ave4 (i, j) or more, then the average photometric values ave (i, j) of the screen G (i, j) is converted using the following equation (12): if ave (i, j) ≥ 3 × ave4 (i, j) then ave (i, j) = ave4 (i, j) + (1/4) x (ave (i, j) - ave4 (i, j)) (12)

When the smoother 2102 The average photometric values ave (i, j) of the screen G (i, j) are calculated as described above. Further, by limiting ave (i, j) to a predetermined lower limit in the limiter 2105 mean photometric values that are too low are prevented from being output. Then, the average photometric value becomes the memory 2103 output. It is noteworthy that in the present invention, depending on the mode of use, it may be possible to output photometric values whose range through the limiter 2105 is limited, without a smoother 2102 provided.

The adjustment circuit 122 for binarization thresholds sets the binarization threshold TH (i, j) based on ave (i, j). The setting unit is a multiplier for multiplying ave (i, j) by a preset coefficient Cb, however, if the coefficient Cb is made equal to x / 16 or Cb is made equal to x / 8 (where x is a predetermined value that is a natural one) Representing number not greater than the denominator), then the adjustment circuit 122 for a binarization threshold simply by an adder that is efficient in both cost and speed.

The binarizer 123 binaries each pixel of the block Block (i, j) corresponding to the screen G (i, j) based on the binarization threshold TH (i, j). The compressor 124 performs the compression suitable for a binary image such as MH or MR. The compressed image is then stored in a frame buffer 125 saved.

Next, the flow of processing will be described as far as the binarization of a multi-valued image is concerned. 22 Fig. 10 is a flowchart showing the flow of image data processing as far as the binarization of a multi-valued image is concerned. First, the CPU reads 108 the size of the image from the CCD 101 (Step S2201). Next, the CPU sets 108 a screen G (i, j) among the photometric values obtained from the photometer 2101 is outputted (step S2202), and calculates the average value of that screen (step S2203). All photometric values, or a suitable sample thereof, can be used to determine the mean.

Next, a determination is made as to whether or not the average of the photometric values of the screen with the means values of the photometric values of adjacent screens is inconsistent (step S2204). If the mean value is inconsistent (ie, the result of the determination in step S2204 is YES), then the average value of the photometric values is calculated using the smoother 2102 smoothed (step S2205). If the mean value is not inconsistent (ie, the result of the determination in step S2204 is NO) or, alternatively, if the smoothing process of step S2205 has been completed, then a determination is made as to whether or not the output photometric values are too low (step S2206). If the output photometric values are too low (ie, the result of the determination in step S2206 is YES), then the photometric values are replaced by preset values (step S2207). If the output photometric values are not too low (ie, if the result of the determination in step S2206 is NO) or, alternatively, if the values have been replaced in step S2207, then a binarization threshold is calculated based on the photometric values of the screen G (i, j) (step S2208).

Next, the block (i, j) corresponding to the screen G (i, j) is extracted from the frame memory 107 read by the block read controller 2104 (Step S2209) and the multivalued image within the block is binarized by using the binarization threshold (Step S2210). A determination is made as to whether or not binarization of all the blocks has been completed (step S2211). If binarization of all the blocks has been completed (ie, if the result of the determination in step S2211 is YES), the binarization processing is ended. If the binarization of all the blocks has not been completed (ie, if the result of the determination in step S2211 is NO), the routine returns to step S2202, the next screen is set (for example, G (i + 1, j)), and the steps 2202 and S2211 repeated.

Because a binarization threshold is set by using information (photometric values) obtained from an automatic photometry section which is provided in a digital camera in this eleventh embodiment no need after a separate processing to set the binarization threshold. As a result, the circuitry is simplified, and the cost it will be reduced. Because a separate processing for calculation a binarization threshold is unnecessary, the power consumption can be reduced. Even if the picture Point reflections from a light source is a binarization high quality a multivalued image in addition possible. By a smoother and a delimiter used in combination will be beyond Binarizing thresholds effectively set.

It is noteworthy that a CMOS sensor 101 in the digital camera 2100 however, depending on the mode of use, a design may be used where a CMOS sensor is used 2301 in a digital camera 2300 is used as it is in 23 is shown. By using this type of structure, frame memories and a block read controller become unnecessary, and the circuit construction becomes compared with the digital camera 2100 easier. The CMOS sensor 2301 can use the same power source as a CMOS integrated circuit, which also allows power consumption to be reduced. As a result, the extent of the circuit of the digital camera 2300 made smaller than the digital camera 2100 , And the power consumption is also smaller. Consequently, it is possible to provide a more efficient digital camera than the digital camera 2100 in terms of power consumption, processing speed and cost.

The twelfth embodiment explained below relates to an image capturing apparatus that uses binarization thresholds to be applied to each pixel within a predetermined block based on photometric values that have been limited to a predetermined range. The image capture apparatus is used in a digital camera. The 24 Fig. 10 is a diagram showing an example of the structure when such an image sensing apparatus is used in a digital camera.

The digital camera 2400 is equipped with: the CCD 101 , the A / D converter 102 , the white balance adjuster 103 , the pixel interpolator 104 , the luminance generator 105 , the aperture corrector 106 , the frame store 107 , the CPU 108 , the photometer 2101 , the smoother 2102 , a limiter 2401 , the setting circuit 1701 for the screen binarization threshold, the memory 1702 , the interpolator 1703 for the binarization threshold, the binarizer 123 , the compressor 124 and the image memory 125 , It is noteworthy that a section for generating color difference signals is omitted.

The CCD 101 converts light collected by an unillustrated digital camera optical system 2400 into electrical signals and outputs analog R, G, B signals for each pixel, forming a multi-valued image. The output analog signals are converted into digital signals by the A / D converter 102 converted. If you go through the white balance adjuster 103 , the pixel interpolator 104 , the luminance generator 105 and the aperture corrector 106 The digital signals are subjected to processing such as luminance value interpolation, extraction, and the like, and then temporarily stored in the frame memory 107 saved.

The CPU 108 controls each of the circuits and links of the digital camera 2400 , The photometer 2101 is equipped with an automatic exposure detection mechanism (A / E detection mechanism) to perform a photometry on an object before an image is photographed, and measures the brightness of each screen based on digital signals received from the A / D converter 102 be issued. The method of photometry may, for example, comprise a measurement by adding the luminance values of the pixels. At this time, the CPU can 108 also sample photometric values that are added by the photometer 2101 to be used.

The smoother 2102 smoothes the photometric values of each screen from the photometer 2101 and gives it as ave (i, j) to the delimiter 2105 out. In the limiter 2401 ave (i, j) is limited by a preset lower limit, so that mean photometric values that are too low are prevented from being output.

The adjustment circuit 1701 for the screen binarization threshold sets the binarization threshold TH (i, j) based on ave (i, j). The thresholds for block binarization for all blocks are in memory 1702 saved. The binarization thresholds for the individual pixels are next passed through the interpolator 1703 for the binarization thresholds using the block binarization thresholds for all blocks. The luminance values for each pixel are acquired simultaneously from the frame memory 1703 while calculating the binarization thresholds for the pixels. The luminance values and the binarization thresholds are put into the binarizer 123 entered. In the binarizer 123 the binarization thresholds and the luminance values are compared and the luminance values are binarized. The binarized image then undergoes image compression in the compressor 124 which are suitable for a binarized image, for example, MH or MMR. The compressed image is then in the image memory 125 saved. It is noteworthy that a digital camera with the same type of structure is achievable even if a CMOS sensor is used instead of the CCD 101 is used.

Next, the flow of processing until a multi-valued image is binarized will be described. The 25 Fig. 10 is a flowchart showing the flow of image data processing until a multi-valued image is binarized. The first thing the CPU reads 108 the size of the image from the CCD 101 (Step S2501). Next, the CPU reads 108 a screen G (i, j), from or below the photometric values obtained by the photometer 2101 is outputted (step S2502) and calculates the average value of that screen (step S2503). The average can be determined by using all photometric values or by using an appropriate sample thereof.

Next, a determination is made as to whether or not the average value of the photometric values of the screen is inconsistent with the average values of the photometric values of the adjacent screens (step S2504). If the mean value is inconsistent (ie, the result of the determination in step S2504 is YES), then the mean value of the photometric values is smoothed by the smoother 2102 is used (step S2505). If the mean value is not inconsistent (ie the result of the determination in step S2504 is NO) or alternatively, if the smoothing process of step S2504 has been completed, then a determination is made as to whether or not the photometric values are too low (step S2506). If the photometric values are too low (ie, if the result of the determination in step S2506 is YES), those photometric values are exchanged with predetermined values (step S2507). If the photometric values are too low (ie, if the result of the determination in step S2506 is NO) or, alternatively, if the replacement process of step S2507 has been completed, the block binarization threshold TH (i, j) of the block corresponding to the screen becomes , calculated based on the photometric values of the screen G (i, j) (step S2508).

A Determination is then made as to whether or not a block binarization threshold exists for all blocks has been calculated (step S2509). If the block binarization thresholds all blocks have been calculated (i.e., if the result of the determination in step S2509 YES), the process goes to the next Step over. If the block binarization thresholds all blocks have not been calculated, (i.e., if the result of the determination is NO in step S2509), then the block which becomes the block G (i, j) is adjacent (for example G (i + 1, j)) fixed and Steps S2502 to S2509 are repeated.

If the block binarization thresholds for all blocks have been calculated (ie, if the result of the determination in step S2509 YES), an interpolation block is set (step S2510). The interpolator 1703 for the binarization threshold, then binarization thresholds th (x, y) of each individual pixel are calculated by interpolation using the binarization thresholds (step S2511). It is noteworthy that x and y are natural numbers representing the position of each pixel in an image. Next are the pixels from the frame memory 107 is read (here assumed to be g (x, y)) by using the above binarization thresholds th (x, y) (step S2512). A determination is made as to whether or not binarization processing for all pixels has been completed (step S2513). If the binarization processing has not been completed for all pixels (ie, if the result of the determination in step S2513 is NO), then a pixel adjacent to the pixel g (x, y) (for example, the pixel g (x + 1, y)), and steps S25011 to S2613 are repeated.

If however, a binarization processing for all pixels completed (that is, if the result of the determination in Step S2513 is YES), a determination is made as to whether or if not a binarization at all interpolation blocks has been completed (step S2514). If a binarization not with all interpolation blocks completed (that is, if the result of the determination in Step S2514 is NO), then steps step S2510 until S2514. If the binarization on all interpolation blocks is completed (i.e., that the result of the determination in step S2514 YES), then the processing is ended.

Because Binarizing thresholds be set by using information (photometric values) which are achieved by an automatic photometric section which is provided with a digital camera is at this twelfth embodiment a separate processing for the setting of binarization thresholds is unnecessary. Consequently For example, the circuitry can be simplified and the cost can be reduced become. Because processing for separately calculating binarization thresholds unnecessary is, the power consumption can also be reduced. Even if point reflections due to a light source in an image are present, is additional a binarization higher quality a multi-valued image possible. By doing a smoother and a delimiter in combination will continue to be used Binarizing thresholds set efficiently. Because binarization thresholds, apply to every single pixel in an interpolation block are, and a binarization based on photometric Values performed is, high quality image processing is possible. If photographic notepads or Notepads are used especially in those cases when For example, every character is small or the characters in a section the area are very small, a Teilbinärisierung high quality the relevant areas possible.

The thirteenth embodiment explained below relates to an image capturing apparatus that performs image binarization by software processing in the CPU. 26 Fig. 10 is a block diagram showing an example of the apparatus construction from the input of an image until a binarized image is recorded when an image capturing apparatus which performs image binarization by software processing in the CPU is used in a digital camera. Notably, in the present embodiment, structural members identical to those previously described in the embodiments are identical, denoted by the same reference numerals, and a detailed description thereof is omitted. The description will be given here mainly for those portions which are different from the previous embodiments.

The digital camera 2600 includes the CCD 101 , the A / D converter 102 , the white balance adjuster 103 , the pixel interpolator 104 , the luminance generator 105 , the aperture corrector 106 , the frame memory 107 , a CPU 2601 , a ROM 2602 , a ram 2603 and the image memory 125 , A section for generating color difference signals will be omitted.

The CCD 101 converts light through an optical system, not shown, of the digital camera 2600 is collected into electrical signals and outputs analog R, G, B signals of each pixel forming the multivalued image. The output analog signals are converted into digital signals by the A / D converter 102 converted. The digital signals are then subjected to a variety of processing, such as interpolation of the luminance values, extraction, and the like as performed by the white balance adjuster 103 , the pixel interpolator 104 , the luminance generator 105 and the aperture corrector 106 and then they are temporarily stored in the frame memory 107 saved.

The CPU 2601 controls each of the circuits and links of the digital camera 2600 and has the function of binarizing the image. A software program for performing the binarization function is in the ROM 2602 saved with the CPU 2601 connected is. The RAM 2603 stores data such as image data and the like and acts as a work area to perform the binarization process.

The software program to perform the functions of calculating the mean luminance, the low luminance threshold setter, the binarization threshold setter, the binarization threshold interpolator, the binarizer, the compressor, and the like described in U.S. Patent Nos. 4,996,231; In the foregoing embodiments, in the ROM 2602 saved. The calculation of the mean luminance, the setting of the binarization threshold, the binarization, the compression of the binarized image and the like becomes in the CPU 2601 done by running the program. An image that has been binarized and that in the CPU 2601 is compressed in the image memory 125 saved. Depending on the mode of use, a software program for performing the functions of the photometer, the smoother, the limiter, the block binarization threshold setting circuit, the binarization threshold interpolator, the binarizer, the compressor and the like may be used in the ROME 2602 in the digital camera 2600 can be stored, and the setting of the binarization threshold, the binarization, the compression of the binarized image and the like can be performed by the program in the CPU 2601 expires. An image that has been binarized and that in the CPU 2601 is compressed in the image memory 125 saved. It is also possible to use a CMOS sensor instead of the CCD 101 to use.

In this thirteenth embodiment, since there is every function by the CPU 2601 and the ROM 2602 No need for a separate section to perform those of the functions, and development costs can be reduced. This allows a digital camera to be provided at a low cost. In addition, updating the software makes it possible to consistently provide a new algorithm.

The fourteenth embodiment explained below relates to an image sensing apparatus equipped with a photometer for performing image binarization in the CPU by software processing. The 27 Fig. 10 is a block diagram showing the apparatus construction from the input of an image to a binarized one when an image capturing apparatus which performs image binarization by software processing in the CPU is used in a digital camera. It is noteworthy that in the present embodiment, a detailed description of the structural elements that are identical to those of the thirteenth embodiment will be omitted. The description given here consists mainly of those parts which are different from the thirteenth embodiment.

The digital camera 2700 includes the CCD 101 , the A / D converter 102 , the white balance adjuster 103 , the pixel interpolator 104 , the luminance generator 105 , the aperture corrector 106 , the frame memory 107 , a CPU 2701 , a ROM 2702 , a ram 2703 , a photometer 2704 , a store 2705 and the image memory 125 , A section to generate color difference signals is omitted.

The CCD 101 converts light through an optical system, not shown, of the digital camera 2700 is collected into electrical signals and outputs analog R, G, B signals for each pixel forming the multi-valued image. The output analog signals are converted into digital signals by the A / D converter 102 converted. The digital signals are then subjected to a variety of processing, such as interpolation of the luminance values, extraction, and the like when passing through the white balance adjuster 103 , the pixel interpolator 104 , the luminance generator 105 and the aperture corrector 106 and then they are temporarily stored in the frame memory 107 saved.

The photometer 2704 is equipped with an automatic exposure detection mechanism (AE detection mechanism) to perform photometry on an object before an image is photographed, and measures the brightness of each screen based on the digital signals received from the A / D converter 102 be issued. The method of photometry represents a measurement by adding the luminance values of the pixels. At this time, it is for the CPU 2701 also possible, the photometric values to be used in the addition, by the photometer 2704 scan. Because it is possible to generate luminance signals at the same time as average luminance values in the photometer 2704 is the amount of calculation in the CPU 2701 reduces, and the processing of the photograph of the image to the storage of the binary image can be performed faster than in the thirteenth embodiment.

In addition to the function of controlling each circuit and each member of the digital camera 2700 becomes the CPU 2701 equipped with an image binarization function. A software program for performing the binarization function is in the ROM 2702 saved that with the CPU 2701 that with the CPU 2701 connected is. The RAM 2703 stores data such as image data and the like and acts as a work area to perform the binarization processing.

A software program for performing the functions of the smoother, the limiter, the block binarization threshold setting circuit, the binarization threshold interpolator, the binarizer, the compressor, and the like may be provided in the ROM 2702 and setting the binarization threshold, binarization, compression of the binarized image and the like can be performed by the program in the CPU 2701 is performed. An image that has been binarized and that in the CPU 2701 is stored in the image memory 125 saved. Depending on the mode of use, it is also possible to use a CMOS sensor instead of the CCD 101 to use.

Because every function by the CPU 2701 and the ROM 2702 is performed in this fourteenth embodiment, in the same manner as in the thirteenth embodiment, there is no need for a separate section to perform each of the functions, and the development cost can be reduced. This enables a digital camera to be provided at a low cost. In addition, updating the software makes it possible to constantly provide new algorithms.

The present invention can also be achieved by using software in addition to the above-described embodiments. 28 shows an example of the computer system that can realize the present invention by using software.

In 28 leads the CPU 2801 the overall control of the apparatus based on a control program; the ROM 2802 saves the control program; 2803 is the RAM; the display device 2804 indicates input states and output states of a computer; 2805 is a hard disk; the keyboard 2806 is used to enter strings and the like; 2807 is a CD-ROM drive. The CD-ROM 2808 serves as a computer-readable storage medium on which a program for executing the image binarization method of the present invention is recorded.

In the computer system having the above structure, a program for executing the image binarization method of the present invention is provided on the CD-ROM 2808 saved. This program is read and controlled by the CPU 2801 enabled, allowing the image binarization process to be performed. The binarized information becomes the hard disk 2805 or the like.

As may be described above the image binarization apparatus according to one aspect The present invention provides low luminance values from an object block based on mean luminance values of the surrounding blocks and binarization thresholds can be set for the object blocks based on the mean of the luminance values, from which the low luminance values have been removed. consequently is a binarization high quality a multi-valued image possible.

at the image binarization apparatus according to one Another aspect of the present invention may be low luminance values are removed, based on mean luminance values of the surrounding blocks, and binarization thresholds can for the Object blocks set based on the mean of the luminance values, from which the low luminance values have been removed. In addition, binarization thresholds, apply to each pixel within an interpolation block are set based on binarization thresholds of adjacent ones Object blocks. Consequently, a binarization high quality a multi-valued image possible.

At the the image binarization apparatus according to one more Another aspect of the present invention may be low luminance values be removed from an object block, based on middle Luminance values of the surrounding blocks, and the average of the luminance values, of which the low luminance values can be rounded to values that are within one predetermined range lie. In addition, the Binarizing thresholds for the object blocks be set based on these values. Consequently, one is binarization high quality a multi-valued image possible.

In the image binarizing apparatus according to still another aspect of the present invention, low-luminance values based on average luminance values of the surrounding blocks can be removed, and the average of the luminance values from which the low luminance values have been removed can be rounded to values within a predetermined range. In addition, binarization thresholds to be applied to each pixel within an interpolation block may be based on binarization thresholds barter object blocks are set. Consequently, high quality binarization of multivalued images is possible.

Further will be the size of the generated Blocks in accordance with the picture size or changed with the total number of pixels of the multivalued image. consequently can be a suitable block size for the size of the characters in accordance with the size or with the total number of pixels of the multivalued image. Consequently, a binarization high quality the multivalued image possible.

Further will the size or the shape of the blocks, which are to be generated in accordance with the positions of the blocks, which are to be generated, changed within the multi-valued image. Consequently can detailed corrections resulting from the optical system such as, for example, peripheral light reduction. consequently is a binarization high quality a multi-valued image possible.

Of the Image Binärisierungsapparat includes a scanning unit that scans pixels representing the multivalued image and the luminance value output unit outputs luminance values of the pixels scanned by the scanning unit, so that the number of pixels, if a luminance value in a block is calculated, can be reduced. Consequently, a binarization high quality a multi-valued image quickly and with a low power consumption carried out become.

Further the scanning unit sets a sampling interval, the sampling in accordance with the image size of the entire Number of pixels and block size used becomes. Even if the block size changes, the number of pixels can be reduced are or are kept constant when mean luminance values calculated within object blocks become. Consequently, binarization can be higher quality a multi-valued image can be performed quickly and with low Power consumption.

The Calculation unit for the average luminance value further comprises an adder unit for generating the Luminance values for to add each pixel, and a counting unit to the number of To count pixels that are added by the adder unit, and when the number of Pixels passing through the counting unit were counted, is a power of two, the adder determines a middle one Luminance value. As a result, a divider is not necessary if a mean we calculated, and a simple structure that using only one adder can be used. Consequently, can a binarization high quality a multivalued image can be done quickly, and a lower one Power consumption can be achieved.

According to one more Another aspect of the present invention may be applied to the image capture apparatus Binarizing thresholds of the generated blocks on the basis of smoothed photometric values of the generated screens. consequently is a binarization high quality a multi-valued image possible.

at the image capture apparatus according to one more another aspect of the present invention, binarization thresholds, which are applied to each pixel of an interpolation block can, on the basis of smoothed photometric values of adjacent generated screens become. Consequently, a binarization high quality a multi-valued image possible.

at the image capture apparatus according to one more Another aspect of the present invention may be the photometric values the umbrellas produced are rounded so that they are within one predetermined range, and the binarization thresholds of the generated blocks can be set based on these values. Consequently, one is binarization high quality a multi-valued image possible.

at the image capture apparatus according to one more Another aspect of the present invention may be the photometric values the adjacent generated screens are rounded so that they within a predetermined range, and the binarization thresholds can be applied to each pixel of the interpolation blocks, based be set to these values. Consequently, a binarization high quality a multi-valued image possible.

The Screens generated by the screen partitioning unit are still identical to blocks, which are generated by the block dividing unit. Consequently Is it possible, the correlation between a photometric value of a screen and a luminance value of a block. Consequently, it is the binarization high quality a multi-valued image possible.

According to one more Another aspect of the present may be in the image binarization process low luminance values of object blocks based on mean Luminance values of surrounding blocks are removed, and binarization thresholds of object blocks can adjusted on the basis of the mean value of the luminance values from which the low luminance values have been removed. Consequently, the binarization is high quality a multi-valued image possible.

at the image binarization process according to one more Another aspect of the present invention may be low luminance values based on average luminance values of the surrounding blocks and binarization thresholds the object blocks can adjusted on the basis of the mean value of the luminance values from which the low luminance values have been removed. additionally can be binarization thresholds, which are to be applied to each pixel within an interpolation block based on binarization thresholds of adjacent ones object blocks be set. Consequently, the high quality binarization is one multivalue image possible.

at the image binarization process according to one more Another aspect of the present invention may be low luminance values of object blocks are removed, based on the average luminance values the surrounding blocks, and the unit of luminance values, of which the low luminance values can be rounded to within a predetermined range Area to lie. additionally can be binarization thresholds based on object blocks be set to these values. Consequently, the binarization is high quality a multi-valued image possible.

at the image binarization process according to one more Another aspect of the present invention may be low luminance values based on mean luminance values of the surrounding blocks and the unit of luminance values, of which low luminance values can be removed be rounded so that they are within a predetermined range lie. additionally can Binarization thresholds of object blocks based on these values, and binarization thresholds that are set can be applied to each pixel of an interpolation block, based on binarization thresholds from adjacent object blocks be set. Consequently, the high quality binarization is one multivalue image possible.

In The block dividing step matches the size of an object block with the image size of the multivalued Image or with the total number of pixels of the multivalued image changed. As a result, it is possible to select a block size that for the Size of the characters in accordance with the picture size or the total number of pixels possible is. As a result, binarization is higher quality a multi-valued image possible.

In The block dividing step becomes the size or shape of the blocks that to be created in accordance with the positions of the blocks, which are to be generated, changed within the multi-valued image. Consequently are detailed correctors, different from the optical system, as they result from the optical system, such as a light reduction in the periphery or on the edge, possible. consequently is a binarization high quality a multi-valued image possible.

Further For example, a scanning step for scanning pixels is provided forming the multivalued image, and in the removal step low luminance values are determined using pixels that were sampled in the sampling step, luminance values below of the low luminance threshold from the luminance values of the pixels and only luminance values that exceed the luminance threshold, are issued. As a result, the number of pixels that can be used is calculated when a mean luminance value within a block will be reduced. Consequently, the high quality binarization of a multi-valued image can be achieved quickly, and the power consumption can be reduced.

Further For example, a scanning step for scanning pixels is provided forms the multi-valued image, and in the middle luminance value calculating step Average luminance values are calculated using pixels that be sampled in the sampling step. As a result, the number the pixel used when calculating a mean luminance value will be reduced in one block. Consequently, the binarization high quality a multi-valued image are achieved quickly and this at a low power consumption.

In the sampling step will continue to be a sampling interval that occurs during sampling is used in accordance with the picture size, the Total number of pixels or block size set or set. As a result, the number of pixels used when a mean luminance value is calculated within a block, be reduced or can be kept constant, and even if the size of the block changes. Consequently, a binarization high quality a multi-valued image can be achieved quickly, and the power consumption can be reduced.

at the image acquisition method according to yet another aspect of the present invention Binarizing thresholds generated blocks on the basis of smoothed photometric values of a generated screen. Consequently, a binarization high quality a multi-valued image possible.

In the image acquisition method according to In yet another aspect of the present invention, binarization thresholds applied to each pixel of an interpolation block are adjusted based on smoothed photometric values of adjacent generated screens. Consequently, high quality binarization of a multi-valued image is possible.

at the image capture method according to yet another aspect The present invention produces photometric values Umbrellas are rounded to values that are within a predetermined range lie, and binarization thresholds of generated blocks are set based on these values. Consequently, one is binarization high quality a multi-valued image possible.

at the image capture method according to yet another aspect In the present invention, photometric values of adjacent, generated umbrellas are rounded to values that are within a predetermined Range, and the binarization thresholds that are be applied to each pixel of the interpolation block set based on these values. Consequently, a binarization high quality a multi-valued image possible.

Further are the screens generated at the screen dividing step be identical to the blocks, which are generated in the block dividing step. Consequently becomes the correlation between the screen photometry values and the block luminance values elevated. Consequently, a binarization high quality a multi-valued image possible.

The computer readable recording medium according to yet another aspect of the present invention It is that a computer based on each step of the Image-binarizing, that was described above works. The invention relates continue a corresponding program.

The present invention is based on the priority-based Japanese applications 11-113761, filed in Japan on April 21, 1999 and 2000-035946, which was filed in Ja pan on 14 February 2000. These registrations are hereby incorporated by reference.

at In a digital camera, a CPU divides a multi-valued image in blocks and choose object blocks which should be the object of processing. A CCD gives Luminance values of the multivalued image. An adjustment device for low luminance thresholds low luminance thresholds based on mean Luminance values of blocks which are adjacent to the object block. A calculation device the average luminance value calculates average luminance values, where Luminance values are used, from which luminance values are removed which do not have the low luminance threshold in one Reach the block. A binarization threshold setting circuit sets a binarization threshold of the block based on average luminance values. A binarizer binarizes then the 'multivalent Image in the block based on the binarization thresholds.

1

101

CCD

102

A / D converter

 103

White balance adjuster

104

pixel interpolator

105

Luminanzgenerator

106

Aperturkorrigierer

107

frame memory

108

CPU

109

block buffer

120

calculator for middle luminance

121

adjuster for low Luminance threshold

122

Binarizing threshold adjustment

123

binarizers

124

compressor

125

image memory

5

501

comparator

 502

Addition result register

503

counter

504

shift register

506

adder

507

gate

508

gate

6

S601

Reading Resolution

 S602

Set sampling

S603

Set Block size and subdivision pattern

S604

Set blocks

S605

Do a sample by

S606

remove low luminance value pixels

S608

Compute Binarizing threshold

S609

binarization

S610

is binarization for all blocks Complete?

7

701

CMOS sensor

102

A / D converter

103

White balance adjuster

104

pixel interpolator

105

Luminanzgenerator

106

Aperturkorrigierer

108

CPU

109

block buffer

120

calculator for middle luminance

121

adjuster for low Luminance threshold

122

Binarizing threshold adjustment

123

binarizers

124

compressor

125

image memory

8th

101

CCD

102

A / D converter

103

White balance adjuster

104

pixel interpolator

105

Luminanzgenerator

106

Aperturkorrigierer

107

frame memory

108

CPU

109

block buffer

120

calculator for middle luminance

121

adjuster for low Luminance threshold

801

Block binarization threshold adjustment

802

Storage

803

Binarizing threshold interpolator

123

binarizers

124

compressor

125

image memory

10

S1001

Reading Resolution

S1002

Set sampling

S1003

Set Block size and subdivision pattern

 S1004

Set blocks

S1005

Do a sample by

 S1006

remove low luminance values

S1007

Compute mean luminance value

 S1008

Compute Block binarization threshold

S1009

save Block binarization thresholds In the storage room

 S1010

were for all blocks Block binarization thresholds calculated?

S1011

Set interpolation block

 S1012

Compute Binarizing threshold

S1013

binarization

 S1014

were all pixels of a binarization subjected?

S1015

Has been binarization with all interpolation blocks carried out?

11

1101

CMOS sensor

102

A / D converter

103

White balance adjuster

104

pixel interpolator

105

Luminanzgenerator

106

Aperturkorrigierer

107

frame memory

 108

CPU

120

calculator for middle luminance

 121

adjuster for low Luminance threshold

801

Block binarization threshold adjustment

 802

Storage

803

Binarizing threshold interpolator

123

binarizers

124

compressor

125

image memory

12

101

CCD

102

A / D converter

103

White balance adjuster

104

pixel interpolator

105

Luminanzgenerator

106

Aperturkorrigierer

107

frame memory

108

CPU

109

block buffer

120

calculator for middle luminance

121

adjuster for low Luminance threshold

1201

limiter

1202

Binarizing threshold adjustment

123

binarizers

124

compressor

125

image memory

13

101

CCD

102

A / D converter

103

White balance adjuster

104

pixel interpolator

105

Luminanzgenerator

106

Aperturkorrigierer

107

frame memory

108

CPU

109

block buffer

120

calculator for middle luminance

121

adjuster for low Luminance threshold

1301

limiter

801

Block binarization threshold adjustment

802

Storage

803

Binarizing threshold interpolator

123

binarizers

124

compressor

125

image memory

14

101

CCD

102

A / D converter

103

White balance adjuster

104

pixel interpolator

105

Luminanzgenerator

106

Aperturkorrigierer

107

frame memory

108

CPU

1401

photometer

1402

smoother

1403

Storage

1404

Block reading control means

122

Binarizing threshold adjustment

123

binarizers

124

compressor

125

image memory

15

S1501

Reading Resolution

 S1502

Set a screen of the photometer

S1503

Compute Photometric values of screens

S1504

are photometric values inconsistent?

S1505

Smooth

S1506

harvest corresponding block from the frame memory

S1507

Compute Binarizing thresholds

 S1508

binarization

S1509

Has been binarization completed by all the blocks?

16

1601

CMOS sensor

102

A / D converter

103

White balance adjuster

 104

pixel interpolator

105

Luminanzgenerator

106

Aperturkorrigierer

108

CPU

1401

photometer

1402

smoother

1403

Storage

122

Binarizing threshold adjustment

123

binarizers

124

compressor

125

image memory

17

101

CCD

102

A / D converter

103

White balance adjuster

104

pixel interpolator

105

Luminanzgenerator

 106

Aperturkorrigierer

107

frame memory

108

CPU

1401

photometer

1402

smoother

1403

Storage

 1404

Block reading control means

1701

Schirmbinärisierungs threshold setting circuit

1702

Storage

1703

Binarizing threshold interpolator

123

binarizers

124

compressor

125

image memory

19

S1901

Reading Resolution

S1902

Set capture a screen of the photometer

S1903

Compute photometric values from the screen

S1904

are photometric values inconsistent?

 S1905

Smooth

S1906

Compute Screen binarizing thresholds

 S1907

save Screen binarization threshold In the storage room

S1908

were Screen binarizing thresholds for all Umbrellas calculated?

 S1909

Set interpolation block

S1910

Compute Binarizing thresholds

S1911

binarization

S1912

Has been for all Pixel the binarization processing completed?

 S1913

Has been binarization for all interpolated block completed?

20

2001

CMOS sensor

102

A / D converter

103

White balance adjuster

104

pixel interpolator

105

Luminanzgenerator

106

Aperturkorrigierer

108

CPU

1401

photometer

1402

smoother

1403

Storage

1701

Screen binarization threshold adjustment

1702

Storage

1703

Binarizing threshold interpolator

123

binarizers

124

compressor

125

image memory

21

101

CCD

102

A / D converter

103

White balance adjuster

104

pixel interpolator

105

Luminanzgenerator

106

Aperturkorrigierer

107

frame memory

108

CPU

2101

photometer

2102

smoother

2105

limiter

2103

Storage

2104

Block read control means

122

Binarizing threshold adjustment

123

binarizers

124

compressor

125

image memory

22

S2201

Reading Resolution

 S2202

Set capture a screen of the photometer

S2203

Compute photometric values of the screen

 S2204

are photometric values inconsistent?

S2205

Smooth

 S2206

are photometric values too low?

S2207

replace photometric values with preset values

S2208

Compute Binarizing threshold

S2209

harvest corresponding block from the frame memory

S2210

binarization

S2211

Has been binarization for all Pixel completed?

23

2301

CMOS sensor

 102

A / D converter

103

White balance adjuster

104

pixel interpolator

105

Luminanzgenerator

106

Aperturkorrigierer

108

CPU

2101

photometer

2102

smoother

2103

Storage

2105

limiter

122

Binarizing threshold adjustment

123

binarizers

124

compressor

125

image memory

24

101

CCD

102

A / D converter

103

White balance adjuster

104

pixel interpolator

105

Luminanzgenerator

106

Aperturkorrigierer

107

frame memory

108

CPU

2101

photometer

2102

smoother

2401

limiter

1701

Blockbinärisierungs threshold setting circuit

1702

Storage

1703

Binarizing threshold interpolator

123

binarizers

124

compressor

125

image memory

25

S2501

Reading Resolution

S2502

Set capture a screen of a photometer

S2503

Compute photometric values from the screen

S2504

are photometric values inconsistent?

S2505

Smooth

 S2506

are photometric values too low?

S2507

replace photometric values with preset values

 S2508

Compute Block binarization thresholds

S2509

were Block binarization thresholds for all blocks calculated?

S2510

Set interpolation block

S2511

Compute Binarizing thresholds

S2512

binarization

S2513

Has been for all Pixel Binarization completed?

 S2514

Has been binarization for all interpolated block completed?

26

 101

CCD

102

A / D converter

103

White balance adjuster

104

pixel interpolator

 105

Luminanzgenerator

106

Aperturkorrigierer

107

frame memory

2601

CPU

 2602

ROME

2603

R.A.M.

 125

image memory

27

101

CCD

102

A / D converter

 103

White balance adjuster

104

pixel interpolator

105

Luminanzgenerator

106

Aperturkorrigierer

107

frame memory

2701

CPU

2702

ROME

2703

R.A.M.

2704

photometer

2705

Storage

125

image memory

28

2801

CPU

2802

ROME

2803

R.A.M.

2804

display device

2805

hard disk

2806

keyboard

2807

CD-ROM drive

2808

CD-ROM

Claims (45)

Image Binärisierungsapparat, comprising: a block partitioning unit, the a multi-valued image in blocks divided; a luminance value output unit, the luminance values for each Outputs pixels representing the multi-valued image; a setting unit for one Binarizing threshold, which is a binarization threshold which is to be used when the multivalued image is binarized becomes; a binarization unit, binarizing the multivalued image based on the binarization threshold; a Adjustment unit for a low luminance threshold, which has a low luminance threshold which is to be used when removing low luminance values become; a block selection unit made up of blocks that pass through the block partitioning unit has been created, which selects blocks to be processed whose multivalued images through the binarization unit to binarize are; a distance unit for low luminance values, the from the luminance value output unit receives the luminance values from those pixels that contain the selected block or the selected ones Form blocks, which removes the luminance values that are lower than the low luminance threshold by the low luminance threshold setting unit has been set and that outputs only those luminance values that the exceed low luminance threshold; and a Calculation unit for a mean luminance value that has a mean luminance value of Luminance values calculated by the distance unit for low Luminance values are output, wherein the adjusting unit for the low luminance thresholds based on the calculated, mean luminance values from a processing block or from processing blocks the one selected Block or the selected one blocks adjusts the low luminance threshold; and the adjustment unit for Binarization thresholds that Binarizing thresholds of the blocks based on the calculated average luminance values. Image binarizing apparatus according to claim 1, wherein the block dividing unit changes the size of the generated blocks in accordance with the picture size of the multi-valued picture or the total number of pixels of the multi-valued picture. Image Binärisierungsapparat according to claim 1, wherein the block dividing unit is the size or Shape of the blocks, in agreement with the positions of the blocks which are to be generated within the multivalued image are, changes. Image Binärisierungsapparat according to claim 1, further comprising a scanning unit, the pixels which form the multi-valued image, wherein the luminance value output unit Outputs luminance values of the pixels sampled by the scanning unit. Image Binärisierungsapparat according to claim 4, wherein the scanning unit is a sampling interval, used in scanning, in accordance with the image size, the Total number of pixels and block size. Image Binärisierungsapparat according to claim 1, wherein the calculation unit for the middle Luminance value includes an adder unit, the luminance values of a of each pixel, and a counting unit which counts the number of pixels passing through the adder unit and the adder unit determines a mean luminance value, if the number of pixels counted by the counting unit represents a power of two or a power of two. Image Binärisierungsapparat, comprising: a block partitioning unit, the a multi-valued image in blocks divided; a luminance value output unit having a luminance value of each pixel forming the multi-valued image; a Binaryization threshold setting unit, which a binarization threshold which is to be used when the multivalued image is binarized becomes; a binarization unit, binarizing the multivalued image based on the binarization threshold; a Adjustment unit for a low luminance threshold, which has a low luminance threshold which is to be used when removing low luminance values become; a block selection unit made up of blocks that pass through the block dividing unit was generated, selecting blocks to be processed; a Distance unit for low luminance values, which from the luminance value output unit, the luminance values of receives those pixels, the selected one Block or the selected Form blocks, which removes the luminance values that are lower than the low ones Luminance threshold set by the setting unit for the low Luminance threshold was set, and those only those luminance values outputs that exceed the low luminance threshold; a Calculation unit for a mean luminance value that has a mean luminance value of Luminance values calculated by the distance unit for low Luminance values were output; a block binarization threshold setting unit, which is a block binarization threshold based on the calculated mean luminance values; and an interpolation block setting unit for setting interpolation blocks cover the pixels that are over two or more adjacent blocks extend wherein the low luminance threshold adjustment unit the low luminance threshold based on the calculated mean Luminance values of a processed block or the processed one blocks sets to the selected one Block or adjacent to the selected blocks are; and the binarization threshold setting unit Binarizing threshold which is to be applied to pixels within the interpolation block are based on the block binarization thresholds the blocks of the Interpolation block. Image Binärisierungsapparat according to claim 7, wherein the block dividing unit generates the size of blocks in accordance with the image size of the multivalued Image or with the total number of pixels of the multi-valued image changes. Image Binärisierungsapparat according to claim 7, wherein the block dividing unit is the size or the shape of the blocks, which are to be generated, in agreement with the positions of the blocks, which are to be generated within the multivalued image changes. Image Binärisierungsapparat according to claim 7, further comprising a scanning unit, the pixels which form the multi-valued image, wherein the luminance value output unit Outputs luminance values of the pixels sampled by the scanning unit. Image Binärisierungsapparat according to claim 10, wherein the sampling unit sets a sampling interval, that when sampling in accordance with the picture size, the Total number of pixels or block size is used. Image binarizing apparatus according to claim 7, wherein the average luminance calculation unit comprises an adder unit that adds the luminance values of each pixel; and a counting unit that counts the number of pixels added by the adding unit, and wherein the adding unit determines a mean luminance value when the number of pixels counted by the counting unit represents a power of two and a power of two, respectively , Image Binärisierungsapparat according to one of the claims 1 to 6, further comprising: a luminance value limiting unit, which limits a range of mean luminance values determined by the Luminance value calculation unit can be calculated so that the values lie within a predetermined range, wherein the binarization threshold setting unit the binarization thresholds based on the blocks on the limited, average luminance values. Image Binärisierungsapparat The claim according to claim 13, wherein the block dividing unit generates the size of blocks in accordance with the image size of the multivalued Image or with the total number of pixels of the multi-valued image changes. Image Binärisierungsapparat according to claim 13, wherein the block dividing unit is the size or the shape of the blocks, which are to be generated, in agreement with the positions of the blocks, which are to be generated within the multivalued image changes. Image Binärisierungsapparat according to claim 13, further comprising a scanning unit scanning pixels, which form the multi-valued image, wherein the luminance value output unit Outputs luminance values of the pixels sampled by the scanning unit. Image Binärisierungsapparat according to claim 16, wherein the sampling unit is a sampling interval, the is used in scanning, in accordance with the image size, the Total number of pixels or block size. Image Binärisierungsapparat according to claim 13, wherein the calculation unit for the middle Luminance value includes an adder unit, the luminance value of a each pixel added; and a counting unit which counts the number of pixels passing through the adder unit and the adder unit has a mean luminance value determines if the number of pixels passing through the counting unit counted were a power of two or a power of two. Image Binärisierungsapparat according to one of the claims 7 to 12, further comprising: a luminance value limiting unit, which limits a range of mean luminance values by the calculation unit for average luminance values were calculated so that the values within a predetermined range lie; where the block binarization thresholds calculated based on the limited mean luminance values become. Image Binärisierungsapparat The claim according to claim 19, wherein the block dividing unit generates the size of blocks in accordance with the image size of the multivalued Image or with the total number of pixels of the multi-valued image changes. Image Binärisierungsapparat according to claim 19, wherein the block dividing unit is the size or the shape of the blocks, which are to be generated, in agreement with the positions of the blocks, which are to be generated within the multivalued image changes. Image Binärisierungsapparat according to claim 19, further comprising a scanning unit scanning pixels, forming the multivalued image, wherein the luminance value output unit Luminance values of the pixels sampled by the scanning unit were. Image Binärisierungsapparat according to claim 22, wherein the sampling unit is a sampling interval, the is used in scanning, in accordance with the image size, the Total number of pixels or block size. Image Binärisierungsapparat according to claim 19, wherein the average luminance value calculating unit an adder unit comprising the luminance values of each pixel added; and a counting unit which counts the number of pixels passing through the adder unit and the adder unit determines a mean luminance value, if the number of pixels determined by the counting unit is a power of two. An image binarization method of performing a binarization processing with a multi-valued image, comprising: a block dividing step of dividing the multi-valued image into blocks; a block selecting step in which blocks to be processed are selected among the blocks generated in the block dividing step; a low luminance threshold setting step, wherein a low luminance threshold to be used when low luminance values are adjusted based on mean luminance values of the processed blocks that are adjacent to the selected blocks; a low luminance value removal step of removing luminance values below the low luminance threshold from the luminance values from the pixels included in the selected blocks, and outputting only luminance values exceeding the luminance value threshold; a mean luminance value calculating step in which luminance values output at the low luminance value removing step are input and average luminance values of the selected blocks are calculated; a binarization threshold setting step of setting binarization thresholds to be used in the binarization process of the selected blocks based on mean luminance values of the selected blocks calculated in the middle luminance value calculating step; and a binarization step in which each pixel within the selected blocks is binarized using the binarization thresholds set in the binarization threshold setting step. Image binarize according to claim 25, wherein in the block dividing step, the size of a chosen Block in accordance with the image size of the multivalued Image or with the total number of pixels of the multivalued image changed becomes. Image binarize according to claim 25, wherein in the block dividing step Size or the shape of the blocks, which are to be generated, in agreement with the positions of the blocks, to be changed within the multi-valued image. Image binarize according to claim 27, further comprising a sampling step for sampling the pixels comprising the multi-valued image and wherein at the removal step for low Luminance values using pixels included in the sampling step were sampled, luminance values below the low luminance threshold of the luminance values of the pixels are removed and only luminance values, that exceed the luminance threshold, be issued. Image binarize according to claim 28, wherein in the sampling step a sampling interval, used in the scanning, in accordance with the image size, the Total number of pixels or block size is set. Image binarize to perform of binarization processes with a multi-valued image, comprising: a block partitioning step, in which the multi-valued image is divided into blocks; one Block selection step, in which blocks, which are to be processed, are selected from among the blocks that are to be processed at the block dividing step be generated; a low luminance threshold setting step, where a low luminance threshold to use is when low luminance values are removed, based on mean Luminance values are set by processed blocks that adjacent to the selected ones blocks are; a removal step for low luminance values the luminance values below the low luminance threshold are removed from the luminance values of the pixels contained in the selected blocks which were selected in the block selection step, and where only luminance values, which exceed the low luminance threshold, are output; one Calculation step for mean luminance values, at the luminance values, in the removal step for low Luminance values were output, entered and mean Luminance values of the selected blocks be calculated; an adjustment step for block binarization thresholds, where a block binarization threshold, which is a binarization threshold is on a selected one Block is applied based on a mean luminance value set in the calculation step for medium Luminance values is calculated; an interpolation block setting step, at the interpolation blocks, cover the pixels, which spread over two or more adjacent ones selected Extend blocks, starting from selected ones Blocks set which were selected in the block selection step; one Binarization threshold setting step, at the binarization thresholds, based on pixels within the interpolation blocks to each block binarization threshold of the two or more selected blocks which are bridged by the interpolation block, set in the interpolation block setting step; and a binarization step, where each pixel within the selected blocks is using binarization thresholds binarizes which is in the binarization threshold setting step were set. An image binarization method according to claim 30, wherein in the block dividing step, the size of a selected block is in accordance with the image size of the multi-valued image or changed with the total number of pixels of the multi-valued image. The image binarization process according to claim 30, wherein in the block dividing step the size or Shape of blocks, which are to be generated, in agreement with the positions of blocks, to be changed within the multi-valued image. Image binarize according to claim 30, further comprising a sampling step for sampling of pixels forming the multi-valued image, and wherein at the removal step for low Luminance values using pixels included in the sampling step were sampled, luminance values below the low luminance threshold are removed from the luminance values of the pixels and only luminance values, that exceed the threshold luminance value, be issued. Image binarize according to claim 33, wherein in the sampling step a sampling interval, used in the scanning, in accordance with the image size, the Total number of pixels or block size is set. Image binarize according to one of the claims 25 to 29, further comprising: a luminance value limiting step, in which a range of mean luminance values determined in the calculation step for medium Luminance values were calculated so limited that the values within of a predetermined range, the binarization threshold based on the limited average luminance values. Image binarize according to claim 35, wherein in the block dividing step the Size of one chosen Block in accordance with the image size of the multivalued Image or with the total number of pixels of the multivalued image changed becomes. Image binarize according to claim 35, wherein in the block dividing step the size or the shape of blocks, which are to be generated, in agreement with the positions of the blocks, which are to be generated are changed within the multi-valued image. Image binarize according to claim 35, wherein the image binarization process continues includes a sampling step to sample pixels that are the multi-valued ones Form image, and wherein in the calculation step for the middle Luminance value mean luminance values are calculated using pixels which is sampled at the sampling step were. Image binarize according to claim 38, wherein in the sampling step a sampling interval, used in the scanning, in accordance with the image size, the Total number of pixels or block size is set. Image binarize according to claim 30 to 34 for performing a binarization processing with a multi-valued image, comprising: a luminance value limiting step, in which a range of mean luminance values determined in the calculation step for medium Luminance values were calculated so that the values are limited within of a predetermined range, the block binarization thresholds adjusted based on the limited mean luminance values becomes. Image binarize according to claim 40, wherein in the block dividing step the Size of one chosen Block in accordance with the image size of the multivalued Image or with the total number of pixels of the multivalued image changed becomes. Image binarize according to claim 40, wherein in the block dividing step the size or Shape of blocks, which are to be generated, in agreement with the positions of blocks, to be changed within the multi-valued image. Image binarize according to claim 40, wherein the image binarization process continues includes a sampling step to sample pixels representing the multivalue Form image, and wherein in the calculation step for medium Luminance values average luminance values are calculated by pixels, which were sampled at the sampling step. Image binarize according to claim 40, wherein in the sampling step a sampling interval, used in the scanning, in accordance with the image size, the Total number of pixels or block size is set. Computer readable medium for storing or program with commands that, when executed by a computer, cause the computer to do an image binarization process perform according to one of the claims 25 to 44.